Impact of Gentrified Rural Landscapes on Community Co-Build Willingness: The Differentiated Mechanisms of Immigrants and Local Villagers

Abstract

Rural gentrification is transforming China’s countryside, yet the ways gentrified landscapes shape community co-build willingness across social groups remain unclear. Guided by the Hierarchy Effects Model (HEM) and Martin Phillips’ four-dimensional view of rural landscapes (material, symbolic, social, and living), this study develops a “landscape–emotion–intention” framework linking spatial–environmental continuity, cultural landscape transition, social interaction embeddedness, and new rural livability to community identity, sense of belonging, and co-build willingness. Based on 50 in-depth interviews in She Village, Nanjing, latent Dirichlet allocation (LDA) is used to extract key themes, which are combined with the four-dimensional framework to construct a 25-item questionnaire; 376 valid responses from immigrants and local villagers are then examined through multi-group structural equation modeling and artificial neural networks for robustness and importance analysis. Results indicate that cultural landscape transition and new rural livability are the main drivers of identity and belonging among immigrants, whereas cultural landscape transition, spatial–environmental continuity, and social interaction embeddedness are more critical for local villagers; in both groups, sense of belonging is the strongest predictor of co-build willingness. The study embeds HEM within gentrified rural settings, operationalizes stakeholder perceptions via an LDA–SEM–ANN pipeline, and proposes differentiated strategies for inclusive rural community building and sustainable governance.

1. Introduction

Across countries, rural gentrification—the reinvestment and cultural revalorization of rural places—has moved to the center of debates on community sustainability and urban–rural relations. Against this broader backdrop, China’s suburban villages offer a timely setting to examine these questions with fine-grained stakeholder evidence. The spillover of urban consumer demand and the developmental needs of rural areas have jointly catalyzed the emergence of gentrified landscapes in the countryside, such as boutique guesthouses, themed cafés, and cultural and creative spaces. These landscapes, serving as visible outcomes of the flow of urban–rural resources and vital vehicles for rural transformation, not only markedly enhance the spatial quality and cultural vitality of rural areas but also bear the strategic mission of driving rural economic revitalization and fostering social integration [1,2]. However, achieving sustainable rural community integration—transitioning from spatial sharing to deeper community cohesion—remains a central challenge in rural community development. This study therefore focuses on the core question: how to fully harness the inherent potential of gentrified rural landscapes to effectively foster emotional bonds (community identity and belonging) between immigrants and local villagers, ultimately cultivating and enhancing their shared commitment to building a new type of rural community. Resolving this question is crucial for gauging the success of the gentrification transformation in rural areas—namely, whether it can transcend mere economic and spatial changes to become a genuine core driver of sustainable and inclusive development in rural societies.

In recent years, the international academic community has increasingly focused on community governance and the cultivation of social capital, particularly the significance of community co-construction willingness for rural sustainable development. This study defines community co-construction willingness as: the strength of intention among immigrants and local villagers to actively participate in collective actions aimed at promoting community integration, enhancing collective well-being, and building sustainable communities, based on their sense of identification and belonging to the community. This indicator emphasizes deep social integration founded upon mutual understanding, emotional attachment, reciprocal support, and a shared vision for the future [3,4]. The strength of this willingness is pivotal to the endogenous momentum and governance performance of community development. During China’s current period of accelerated rural opening-up and transformation, fostering a shared commitment to co-construction is a crucial linchpin for ensuring the gentrification process benefits diverse stakeholders and achieves sustainable, endogenous rural development. By adopting community co-construction willingness as a core evaluation metric, this study effectively reflects the genuine perceptions of local villagers and immigrants regarding the spatial, cultural, social, and economic dimensions of rural gentrification.

A series of Chinese policies, exemplified by national-level strategies such as the rural revitalization initiative [5,6,7], have laid a solid policy foundation for the development of gentrified landscapes. Practices have demonstrated that such landscapes contribute to stimulating rural economic vitality and cultural renewal [8,9,10]. However, existing academic research has primarily focused on one-dimensional analyses, such as landscape production under rural gentrification [11], the emotions of immigrants and local villagers [12], and behavioral intentions [13]. Despite its significant integrative value, academia has yet to provide a systematic explanation of the specific psychosocial mechanisms and group-differentiated pathways through which gentrified landscapes empower community development. Consequently, this study proposes a theoretical framework of ‘landscape perception-emotional connection-co-construction willingness’ to explore key factors influencing the willingness of immigrants and local villagers to engage in community co-construction. It seeks to address the following questions: Firstly, how do the multidimensional elements of gentrified landscapes differentially trigger emotional responses among immigrants and local villagers? Second, how do these emotional responses subsequently shape their willingness to engage in community co-creation? Finally, are there significant differences in the factors influencing co-creation willingness between immigrants and local villagers? By examining the factors influencing gentrified landscapes within community-building processes, this research aims to provide more scientifically grounded theoretical guidance and practical support for rural community development, thereby contributing insights and solutions for sustainable rural development.

In summary, the contributions of this study are mainly reflected in the following three aspects: (1) Unlike previous studies highlighting perspectives such as health and tourism, this research shifts scope to the impact of gentrified landscapes on rural community development, thereby enriching the research perspectives on rural gentrification. (2) Through grouping and analysis of data on immigrants and local villagers, this study analyses the different impact paths on different groups and proposes design guidelines based on the results, providing a practical basis for targeted policies to resolve the core contradictions in the process of rural gentrification. (3) This study uses latent Dirichlet allocation (LDA) theme modeling to extract more authentic and comprehensive information on perceptions of gentrified landscapes from offline interviews. In addition, by combining structural equation modeling (SEM) with artificial neural network (ANN) models, this study not only verified the impact of multi-dimensional factors of gentrified landscapes on community co-build willingness, but also revealed the complex non-linear relationships between different variables.

2. Literature Review and Hypotheses

2.1. Rural Gentrification and Gentrified Landscapes

Gentrification was first coined by British sociologist Ruth Glass in 1964 to describe the phenomenon of the urban middle class moving into working-class neighborhoods, leading to rising property prices and the displacement of indigenous residents [14]. Rural gentrification, as a variant form of gentrification, introduced by British geographer Martin Phillips, offers a novel spatial analytical perspective and explores gentrification processes within rural contexts. This shift in focus from urban to rural settings in gentrification research reveals both the universal logic of gentrification and the particularities of its rural manifestation [15]. Researchers emphasize the compilation of the value of agricultural or non-agricultural land and its repositioning as middle-class residential land [16]. In the academic context of exploring contemporary rural transformation in China, the phenomenon of gentrification has gradually transcended the traditional critical framework of ‘displacement,’ presenting a unique localized path. In this study, the term “gentrified landscape” specifically denotes composite spatial entities formed by the injection of new immigrant capital driven by urban consumer demand.

Rural gentrification brings about changes in material space practices, spatial representations, and representational spaces in rural areas, and these changes are intertwined and interpenetrate each other [15]. The understanding of gentrified landscapes is further categorized into four principal dimensions: material landscape; social landscape; symbolic landscape; and living landscape [17]. Moreover, Chen et al. note that gentrified landscapes are frequently generated by the emerging middle classes and “rural tourism entrepreneurs” through meticulous renovations of old dwellings and public spaces [11]. Their aesthetic simultaneously embodies local and global characteristics, allowing symbolic landscapes and living environments to permeate one another. This process reshapes local identity and everyday usage through a reconstructed authenticity. Åberg’s research indicates that gentrification is not merely a physical transformation, but rather a ‘perceptual rural landscape’ reproduced through visual means. Its visual narrative reflects the social imagination and future orientation of rural renewal [18]. Sutherland’s research further underscores that the gentrified rural landscape emerges from the interplay between humans and beyond-human elements—such as trees, topography, and wildlife. Settlers co-construct an “experiential landscape” imbued with local sentiment through sensory engagement and habitat shaping, while maintaining a delicate equilibrium between ecological conservation and consumption [19]. Evidently, the gentrification of rural areas has evolved from its initial phases of spatial occupation and social stratification into a multifaceted practice encompassing ecology, aesthetics, and daily life. Its multidimensional landscape logic offers insights for understanding how China’s rural landscapes can achieve integrated development.

China’s unique system of collective land ownership provides the institutional foundation for this integrated development. Unlike the eviction mechanisms under Western private property systems, China’s rural land system has created unique relational ties: first, under China’s unique rural land ownership structure, houses and rural land cannot be sold by individuals, and local villagers retain ownership of their homestead sites, establishing primary economic ties with new immigrants through property leases. Local villagers are not the typical victims displaced by gentrification [20,21]. Second, new immigrants rely on operating rights to develop businesses, which in turn feeds back into the village’s business environment. This symbiotic relationship based on land use rights has enabled the gentrified landscape in places such as Zhejiang to avoid the exclusion of indigenous people and lay the economic foundation for the construction of a deep-rooted community [22].

2.2. Rural Community Building

Rural community development constitutes a significant topic within the field of rural governance and public policy research. The concept of community (Gemeinschaft) was first proposed by German sociologist Ferdinand Tönnies in his 1887 work Gemeinschaft and Gesellschaft [23]. This concept is essentially based on an organic whole formed by blood ties, geographical ties, and spiritual identity. Its core lies in the emotional bonds that form spontaneously among members, a shared value system, and mutual moral responsibilities [24,25]. However, in the face of increased mobility and heterogeneity in contemporary rural areas—the continuous influx of new immigrants, the penetration of urban consumer culture, and the collision of diverse values—the limitations of the Tönnies paradigm have become increasingly apparent. The closed boundaries and homogeneous membership structure assumed by his theory make it difficult to explain how different groups in gentrified villages reconstruct their communities through negotiation.

Academic discussions on rural community development primarily focus on the intrinsic relationship between community participation, social capital, and community resilience. Research indicates that the level of community participation directly influences governance performance and social cohesion. For instance, Jin and Guo constructed a conceptual model of community participation among middle-aged and elderly residents, highlighting functional needs and public facility conditions as key factors promoting their involvement [26]. While the establishment of interpersonal networks serves as a crucial social mechanism for stimulating participation. Tong et al., using urbanized villages in Shenzhen as case studies, revealed the pivotal role of village collectives in land development and community self-organization [27]. They highlighted the significant value of social capital in mobilizing villagers, integrating resources, and compensating for gaps in government services. Feng et al. proposed an exchange ecology model, positing that the material conditions of public spaces, social interactions, and psychological identification intertwine to collectively foster community participation and intergenerational integration [28].

In exploring pathways for rural community development, scholars have proposed diversified development models. Yu et al., adopting a perspective of rural contraction and settlement evolution, constructed a theoretical model for rural settlement development [29]. They emphasized that different regions must select differentiated development models based on resource endowments and spatial patterns to achieve the optimization and restructuring of community functions. Shen et al. found through research on traditional settlements in Quanzhou that integrating cultural heritage preservation with knowledge-based economic transformation can strengthen the reconstruction of social capital and cultural identity, thereby enhancing community resilience and self-organization capabilities [30]. Liang and Peng identified the key success factors for autonomous rural landscape development, emphasizing that the collaborative efforts of community cadres, artists and specialists are crucial for achieving dual improvements in landscape quality and community self-governance [31]. Overall, these studies collectively point to a trend: rural community development is shifting from singular physical transformation towards the construction of integrated systems, emphasizing the formation of a virtuous cycle between governance, cultural renewal and spatial regeneration.

2.3. Co-Build Willingness and Its Differences Among Groups

The willingness to engage in community co-construction serves as a crucial indicator for gauging the vitality and governance capacity of rural societies. In this study, community co-build willingness refers to the subjective inclination of villagers and immigrants to persistently invest resources and share responsibilities in areas such as environmental enhancement, cultural production, and public affairs governance. This concept emphasizes collaborative participation in the long-term development of the community, rather than merely attitudes towards participation in one-off events. Existing research indicates that the formation of co-build willingness is often influenced by multidimensional factors. Zheng et al. suggest that the built environment of a community can impact residents’ health and motivation through social participation and outdoor activities, thereby further enhancing individuals’ motivation to engage in public affairs [32]. Xu et al. found that the frequency of social interaction and the quality of the local social environment significantly influence residents’ levels of community trust, particularly where structural disparities exist between local residents and migrant populations. Overall, the formation of co-build willingness is influenced by a combination of spatial, social and institutional factors.

As rural gentrification and population mobility intensify, disparities among different groups in terms of social capital, spatial accessibility, and economic embeddedness directly influence the manner and depth of their participation in community co-building. Empirical analysis by Lengerer and Franziska demonstrates that an individual’s socioeconomic status, gender, and residential history collectively shape their participation in local politics, associations, and informal community activities. The pressure to ‘assume expected responsibilities’ reinforces the normative nature of participation to some extent. While it also rendering the spectrum between ‘participation/non-participation’ more complex across different groups [33]. Zheng et al. found that the built environment and social participation exert markedly different effects across income groups, with disadvantaged or mobile populations demonstrating a greater propensity for active engagement when provided with favorable environmental support [32].

2.4. Research Hypothesis

In order to systematically analyze the contribution of landscape to community building, this study adopts Phillips’ four-dimensional analytical framework: material landscape; social landscape; symbolic landscape; and living landscape [17]. This study has been adapted to local conditions: spatial–environmental continuity is the core manifestation of the material landscape dimension, focusing on the harmony between the texture of newly built spaces and the traditional appearance of settlements. The cultural landscape transition corresponds to the symbolic landscape dimension, which measures the ability of gentrified spaces to adapt local culture. Social interaction embeddedness maps the social landscape dimension and observes the effectiveness of landscape spaces in catalyzing cross-group interaction. New rural livability focuses on the dimension of living landscape and examines the inclusive improvement of daily life brought about by gentrified landscape.

This paper proposes a theoretical framework of ‘landscape perception—resident sentiment—behavioral intention’ based on HEM. The Hierarchy of Effects Model details the entire process by which an individual forms reactions such as attitudes, intentions, or expectations toward a particular product or object. This theory posits that individuals, after being exposed to stimuli from the external environment and conditions, undergo three consecutive stages of psychological reactions, including cognition, emotion, and intention. Based on their prior experience and knowledge, individuals process information from external stimuli to form cognition, which in turn generates specific emotional reactions, ultimately leading to corresponding behavioral intentions [34]. Thus, when immigrants and local villagers find themselves within the social context of a rural gentrified landscape, the positive and negative impacts they perceive constitute individual perceptions, evaluations, and preliminary judgements of information. On this basis, immigrants and local villagers will generate one or more specific emotional experiences, which in turn further trigger planned and directed intentions among immigrants and local villagers.

2.4.1. Spatial–Environmental Continuity

Spatial–environmental continuity, as a core dimension of the gentrified landscape, denotes the degree to which newly constructed or renovated landscape spaces harmonize, integrate, and maintain continuity with the indigenous natural features, traditional settlement patterns, and architectural fabric of the countryside [17]. It not only influences users’ perception of space, enabling residents to experience the vitality of new spaces while still recognizing the historical imprint and environmental characteristics of the countryside, but also shapes individuals’ emotional responses to the environment.

Research indicates that community landscapes constitute a significant component of community identity for some participants [35]. Derk Jan Stobbelaar and Bas Pedroli note that people can develop a sense of belonging to specific landscapes, typically in areas where they spent their youth or places that witnessed pivotal periods in their lives [36]. Following their study of communities within mountain landscapes, Igor Knez and Ingegärd Eliasson observed that the stronger residents’ attachment/intimacy/sense of belonging to their favorite landscape sites, the greater the happiness they experienced when visiting these places [37].

When spatial environments maintain visual consistency and historical continuity, they evoke residents’ cognitive familiarity and emotional attachment, thereby strengthening environmental belonging and community identity. Therefore, this study proposes the following hypothesis:

H1a: 

The spatial–environmental continuity of gentrified landscapes has a positive impact on community identity.

H1b: 

The spatial–environmental continuity of gentrified landscapes has a positive effect on their sense of belonging.

2.4.2. Cultural Landscape Transition

The cultural landscape transition manifests as the capacity of gentrified landscapes to translate and reinterpret local traditional cultural symbols, serving as a pivotal medium for activating cultural memory and constructing local significance. Through contemporary expressions of traditional symbols, landscapes can evoke emotional resonance and cultural identity among residents.

Existing research indicates that the social nature and intangible attributes of cultural landscape constitute core elements in place-making [38]. Xueyu C & Zhenting L note that perceptions of rural cultural memory are frequently formed through contextual participation [39]. Cocks, Vetter, & Wiersum note that cultural landscapes can serve as conduits for multiple connections between people, place and identity [40]. Boonzaaier and Wel, suggest that cultural landscapes can foster a sense of belonging through cultural naming and narrative construction [41].

According to the aforementioned research, it is evident that cultural landscapes, by imbuing modern spaces with local cultural connotations, possess a cultural landscape transition. The more pronounced this characteristic, the more effectively it can evoke collective memories and emotional resonance with tradition among both immigrants and local villagers, thereby enhancing their cultural identification with and emotional attachment to the community. Consequently, this study proposes the following hypothesis:

H2a: 

The cultural landscape transition of gentrified landscapes has a positive impact on community identity.

H2b: 

The cultural landscape transition of gentrified landscapes has a positive effect on their sense of belonging.

2.4.3. Social Interaction Embeddedness

Social interaction embeddedness refers to the quality and depth of engagement between immigrants and local villagers within the gentrified landscape, encompassing social interactions, shared use of public spaces, and participation in governance. It reflects the degree of social capital integration and relational network reconfiguration, fostering a sense of community through the intersecting restructuring of relational networks.

Existing research indicates that when a locality simultaneously provides tangible and intangible benefits—such as security, familiarity, social connections, and livelihood opportunities—it can foster a sense of place and belonging [42]. Lang & Fink show that the commercial activities of rural social entrepreneurs, through innovative organizational forms, promote social inclusion and local solidarity, addressing social challenges such as unemployment, poverty, social exclusion and marginalization [43], indicate that different types of social communication channels exert varying effects on indicators of social cohesion. The “anchoring channel”, which supports routine communication and cooperation between groups, has been demonstrated to correlate significantly with higher levels of social cohesion and place attachment [44].

Therefore, based upon the above research, landscape spaces provide scenarios for cross-group interaction and collaborative mechanisms, fostering reciprocal trust and social solidarity between immigrants and local villagers. This enhances the social interaction embeddedness, thereby strengthening community identity and the need for belonging. Consequently, this study proposes the following hypothesis:

H3a: 

The social interaction embeddedness that was brought by gentrified landscapes has a positive impact on community identity.

H3b: 

The social interaction embeddedness that was brought by gentrified landscapes have a positive effect on their sense of belonging.

2.4.4. New Rural Livability

The new rural livability serves as a barometer for the extent to which the gentrified landscape fulfills the daily needs of both immigrants and local residents through innovations in production and living functions. This constitutes a vital practical foundation for reconstructing rural value recognition. It fosters emotional attachment by enhancing both the convenience of daily life and the sense of meaning derived from it.

Research indicates that, gentrified landscapes provide diversified residential and lifestyle services through spatial transformation, with some immigrants developing emotional attachments through lived experience [45]. Z. Wu & Ma point out that the gentrification of spaces and the evolution of diverse lifestyle services influence residents’ sense of identity [46]. Xue et al. stated that the material changes brought about by tourism development can influence residents’ rural identity [47].

Based on the aforementioned research, when a gentrified landscape simultaneously provides functional convenience and spiritually fulfilling living scenarios, it can strengthen residents’ utilitarian dependence and emotional connection, thereby fostering community identity and a sense of belonging. Therefore, this study proposes the following hypothesis:

H4a: 

The new rural livability that was brought by gentrified landscapes has a positive impact on rural community identity.

H4b: 

The new rural livability that was brought by gentrified landscapes has a positive impact on sense of belonging.

2.4.5. Community Identity

A shared sense of community identity can strengthen ties and solidarity among villagers and form close social networks. In the process of community building, villagers and immigrants are more likely to reach consensus and form a collective force because of their shared identity with the village.

W. Wu et al. pointed out that acceptance by others, community identity, and life satisfaction may influence each other and form a virtuous cycle over time [48]. Strong group identity can mitigate the adverse effects of social exclusion and significantly enhance social participation [49].

When villagers internalize their communal identity, emotional bonds and mutual trust strengthen, transforming social coexistence into proactive collective collaboration.

Accordingly, this study proposes the following hypothesis:

H5: 

The community identity of immigrants and local villagers has a positive effect on the community co-build willingness.

2.4.6. Sense of Belonging

When villagers have a strong sense of belonging to their rural community, their emotional attachment to it also increases. This emotional bond motivates them to contribute their efforts to the development and improvement of their rural community and actively participate in community-building activities.

Research by Menconi et al. indicates that a sense of belonging can serve as a key performance indicator for measuring community engagement processes in rural areas [50]. Research by Levasseur et al. points out that a greater sense of community belonging further enhances social participation [51].

When residents become deeply immersed in the rural environment on an emotional and symbolic level, a sense of belonging transforms individual sentiment into collective commitment, thereby driving proactive community-building practices. Based on the above research, this study proposes the following hypothesis:

H6: 

The sense of belonging of immigrants and local villagers has a positive effect on the community co-build willingness.

On the basis above, immigrants and local villagers will generate one or more specific emotional experiences, which will further trigger planned and targeted intentions among immigrants and local villagers. This section proposes a research hypothesis with the perception of gentrified landscapes by immigrants and local villagers as the core influencing factor. The research model for constructing the co-build willingness of immigrants and local villagers is shown in Figure 1, which illustrates all the paths of the hypotheses.

3. Materials and Methods

3.1. Study Area

Shecun Village, located in Dongshan Subdistrict, Jiangning District, Nanjing City, Jiangsu Province, is an ancient settlement with over 600 years of history, renowned as the “Premier Ancient Style Village of Jinling”. The village spans approximately 16.46 square kilometers, with a construction area of 167 hectares, cultivated land of 100 hectares, water bodies covering 133 hectares, and forested land extending over 1267 hectares. Characterized by undulating hills and abundant water systems, the village exhibits harmonious spatial organization and rich natural resources. The settlement preserves Ming and Qing architectural complexes, including the “Pan Clan Ancestral Hall,” alongside seven emblematic cultural elements such as ancient wells, trees, and kilns. Since 2017, under the guidance of Jiangning District’s Characteristic Rural Revitalization Plan, Shecun Village has adopted tourism-driven development as its core strategy, promoting integrated growth across agriculture, culture, and tourism sectors. From 15 February 2023, to 1 March 2025, this study conducted over 60 days of fieldwork in She Village, collecting extensive primary textual data.

The dataset comprises 50 in-depth interview transcripts from immigrants and local villagers, totaling 82,257 words. Among these, 27 were from immigrants and 23 from local villagers. The migrant sample encompassed homestay operations (10), catering and café operations (8), cultural and creative entrepreneurship (4), and other service activities (5). All interviews were verbatim transcribed and underwent text preprocessing prior to thematic analysis. To avoid category bias, texts were randomly sampled in equal proportions from both groups during analysis. Records deemed excessively brief or irrelevant to the themes were excluded, thereby ensuring the representativeness and comparability of the perceptual data.

3.2. Data Analysis

First, the study began with data collection, primarily using latent Dirichlet allocation (LDA) topic modeling techniques to process and analyze in-depth interviews and feedback data from immigrants and local villagers. LDA (Latent Dirichlet Allocation) is an unsupervised probabilistic generative model used to automatically discover latent topic structures in text corpora. Its core assumption is that documents are composed of multiple topics, and each topic is represented as a probability distribution of a set of related words [52]. The data was collected through in-depth interviews conducted by the research team, including the real experiences and feelings of immigrants and local villagers in the gentrified village.

Then, the study employs SEM to explore the relationships among these variables and uses path analysis to identify and quantify the impact of different characteristics on community co-construction intentions. SEM is a widely used statistical method in data analysis. It combines techniques such as factor analysis, regression analysis, and path analysis to establish, estimate, and test complex relationships between variables, including causal relationships and the mechanisms of latent variables [53].

Finally, an Artificial Neural Network (ANN) model was introduced as a complementary tool to validate and supplement the structural relationships identified by the SEM. The ANN is a computational framework composed of an input layer, one or more hidden layers, and an output layer, allowing flexible, non-parametric mapping between variables. In this study, the ANN analysis was used to verify the robustness of the SEM results by evaluating the relative importance and directional consistency of key predictors, thereby enhancing the reliability and interpretability of the overall model.

The specific research process is illustrated in Figure 2.

This study was reviewed and approved by the Scientific Research Review Board of Nanjing Forestry University. The research involved non-interventional social-science methods (questionnaires and interviews) and did not include any medical procedures or identifiable personal data. All procedures were conducted in accordance with the Declaration of Helsinki [54]. Verbal informed consent was obtained from all subjects involved in the study. Before each interview or questionnaire, the researcher read a standardized consent script explaining the purpose of the study, voluntary participation, and confidentiality principles.

3.3. LDA Topic Modeling Analysis

To better perform topic clustering analysis, a series of data preprocessing steps are required: (1) Manually remove invalid information, such as emoticons, symbol tags, and blank fields; (2) Use the third-party open-source Python 3.8 (Python Software Foundation, Wilmington, DE, USA) toolkit jieba to perform Chinese word segmentation, converting text data into discrete words to provide a basis for subsequent text analysis and feature extraction; (3) Perform stop word processing to remove frequently occurring but semantically meaningless words from the text, including conjunctions, particles, prepositions, etc. Additionally, a custom dictionary was created containing colloquial terms such as ‘routine’ and ‘village expert.’ The aforementioned data preprocessing methods can effectively enhance the effectiveness and accuracy of thematic clustering.

C_V (Coefficient of Variation) coherence score is calculated by computing the pointwise mutual information values of all word pairs within a topic and normalizing these values to assess the semantic consistency within a topic. A higher score indicates stronger relevance between words within the topic, resulting in greater consistency and interpretability of the topic [52]. As shown in Figure 3, the C_V consistency fluctuates with the increase in the number of topics, generally showing an upward trend. The Dirichlet allocation parameters are set to α = 1 and β = 0.01. When the number of topics is 18, the C_V consistency reaches its maximum value.

Following the initial thematic distribution obtained through LDA topic modeling, this study engaged eight experts with backgrounds in landscape planning and environmental design, each possessing at least five years of relevant professional experience, to review and name the topics. This process enhanced the consistency and readability of the thematic interpretations. Guided by principles of semantic consistency and representativeness, the experts categorized and debated high-probability vocabulary. Ultimately, by synthesizing LDA outputs, existing literature, and the research context, they identified 18 key themes alongside 10 representative keywords. As shown in

Table 1. Topic naming results of LDA.

Perceptual Dimensions	Including Dimensions	Topic Name	Theme Keywords (English Version)
Spatial–environmental continuity	Topic 4	Environmental aesthetics enhancement	Environment, space, beauty, village, main, feature, guests, courtyard, been there, two years
	Topic 6	Reconstruction of Rest Areas	Business, young people, hours, tourists, space, no, children, B&B, village, products
	Topic 14	Ecological landscape restoration	Space, tourists, courtyard, rest, nature, doorway, Double Ninth Festival, construction, B&B, elderly people
	Topic 15	Farmland Landscape Experience	Utilize, cultivate, environment, assist, entrance, villagers, construction, main, several, dance troupe
	Topic 16	Children’s play area layout	villagers, village, food and drink, farming, children, aesthetic sense, what are you doing, rice fields, perspective, photography
Cultural landscape transition	Topic 7	Cultural symbol dissemination	Tourists, role, environment, viral popularity, impact, opinion, reason, choice, merchants, whether
	Topic 8	Festive activities continuing	ordinary people, every day, support, carry out, unable to, real, ability, square, Chongyang Festival, every year
	Topic 12	Space function conversion	Impression, guest, tourist, every day, location, use, B&B, school, small, substitute
	Topic 17	Traditional business model renewal	Issues, B&B, Children, Business model, Tradition, Space, Cultural tourism, Business, Children, Catering
Social interaction embeddedness	Topic 1	Intergroup interaction	Tourists, less than, restaurants, ordinary people, contact, B&Bs, many, children, between, experience
	Topic 5	Youth Entrepreneurship Development Activities	B&B, experience, hope, young people, service, outside world, sometimes, organize, others, participate
	Topic 9	Collaboration among diverse groups	Construction, tourists, B&Bs, operational standards, villages, characteristics, groups, infrastructure, residents, experience
	Topic 13	Public affairs collaboration	B&B, Help, Work, Services, Government, Operational Level, Construction, Outdoor, Research, Stories
	Topic 18	Neighborhood relations	B&B, questions, characteristics, normal, villagers, impressions, hopes, opinions, village, common people
New rural livability	Topic 2	Life services adaptation	Villagers, tourists, face, need, consumption, issues, operational level, investment, village, daily
	Topic 3	Optimization of residential facilities	B&B, environment, problems, hopes, desires, facilities, ordinary people, very large, opening a shop, three years
	Topic 10	Business Policy Support	Environment, publicity, preliminary work, local residents, guests, operations, renovation, assistance, children, updates
	Topic 11	Rural innovation and employment attraction	B&B, space, attraction, issues, characteristics, villagers, young people, entrepreneurship, tourists, interviews

. This procedure drew upon established research practices combining LDA with expert review to enhance the systematic rigor and persuasiveness of the findings [55,56].

3.4. Questionnaire Design

Based on the results of LDA thematic modeling, and drawing upon representative research findings concerning rural gentrification and community integration, the perceived themes extracted were translated into measurable questionnaire items within the context of field research in Jiangsu’s rural communities. The questionnaire items were systematically constructed and revised, as shown in

Table 2. Questionnaire design.

Variable Dimension	Measurement Items	Indicator Description	Reference
Spatial–environmental continuity	Diversified environmental features	Measure respondents’ perception of the visual harmony between traditional and modern landscapes.	Balta, Sıla, and Meryem Atik (2022) [57] Fan, Ding et al. (2025) [58] Wang, Lei et al. (2025) [59]
	Composite spatial functions	Representing the adaptability of public space renovations to accommodate different needs
	Social service levels	Measuring the dual capacity of public service facilities to meet basic and quality needs
Cultural landscape transition	Cultural symbol translation	Measuring the acceptance of traditional cultural elements transformed into modern consumer symbols	Zhao, Jiexiang and Zhu, Jiangang (2025) [60] Li, Peiyuan, and Wencui Zhang (2025) [61] Chen, Huiling, and Wei Tao (2017) [62]
	Consumption scenario iteration	Measuring the innovative appeal of the transition from traditional business models to cultural and tourism experience-oriented scenarios.
	Dynamic Revitalization of Traditional Festivals	Measuring the level of participation and sense of belonging in traditional festive activities within communities
Social interaction embeddedness	Interaction between cross-groups	Reflecting the intensity of interaction between immigrants and villagers in a gentrified landscape	Liu, Jun et al. (2023) [63] Fang, Tingting et al. (2024) [64] Wang, Xinrui, Dandan Huang, and Meiling Wu (2026) [65]
	Immigrant Community Development	Measuring the depth of integration of new immigrant groups into local social networks
	Neighbor Relations Adjustment	Measuring the level of neighborhood adaptability of new and old residents caused by spatial changes
New rural livability	Inclusiveness of consumer services	Reflecting the accuracy of the match between consumer services and residents’ needs	Pang, Yuxin, Wenxin Zhang, and Huaxiong Jiang (2024) [66] Li, Yurui et al. (2020) [67]
	Optimization of residential facilities	Measuring the degree of satisfaction with the compatibility of modern facilities with the local environment
	Rural Innovation and Employment Attraction	Measuring the strength of the attraction of the gentrified landscape to local talent
Community identity	Cultural subjectivity perception	Reflecting resident’s shared sense of belonging to the core values of community culture	Ma, Xiaolong, Yiyuan Zhao, and Weifeng Su (2025) [68] Guan, Jingjing et al. (2025) [69] Zhang, Yingchun (2024) [70]
	Consensus on community values	Assess the strength of identification with core community values
	Collective dignity maintenance tendency	Detect behavioral intentions to actively maintain community image and collective dignity
Sense of belonging	Emotional anchoring depth	Measuring the strength of residents’ sense of belonging to a specific community space	Liu, Zhen et al. (2019) [71] Liao, Liao et al. (2025) [72] Chen, Peipei, Min Zhang, and Ying Wang (2023) [11]
	Perception of community acceptance	Assessing individuals’ sense of recognition and inclusion in community social networks
	Long-term attachment	Assess residents’ psychological tendency to view the community as a long-term place of residence.
Community co-build willingness	Tendency to environmental creation	Reflecting residents’ initiative in environmental creation	Long, Ye, Luan Chen, and Xun Li (2025) [73] Yao, Yuting, Shenghua Lu, and Hui Wang (2021) [74]
	Cultural capital investment	Representative level of resource support for local cultural innovation
	Public affairs involvement	Measuring the intensity of residents’ initiative in community public decision-making and affairs management

.

The questionnaire used in this study is divided into two distinct sections. The first part focuses on obtaining basic demographic and personal information from immigrants and local villagers. The subsequent part of the questionnaire aims to capture respondents’ views and attitudes using a 5-point Likert scale, where ‘1’ indicates ‘strongly disagree’, ‘2’ indicates ‘disagree’, ‘3’ indicates ‘neutral’, ‘4’ indicates ‘agree’, and ‘5’ indicates ‘strongly agree’. The second part included immigrants’ and local villagers’ perceptions of the spatial–environmental continuity of the gentrified landscape, cultural landscape transition, social interaction embeddedness, new rural livability, community identity, sense of belonging, and co-build willingness.

3.5. Data Collection

This questionnaire survey was conducted between 15 February 2023 and 1 March 2025 through a combination of on-site and online methods, and questionnaires were distributed to immigrants (Group A) and local villagers (Group B), respectively. A total of 376 valid questionnaires were collected. The aim was to obtain data on the perceptions of immigrants and local villagers towards the gentrified landscape. The questionnaire design covered seven main variables and included 25 questions. The proportion of villagers and immigrants is shown in

Table 3. Descriptive statistics of respondents’ characteristics.

Demographic Variable	Categories	Frequency	Percentage (%)
Villagers’ household registration	Local	231	61.4%
	Non-local	145	38.6%
Age	16–25 age	76	20.2%
	25–35 age	90	23.9%
	35–50 age	128	34.1%
	50–75 age	82	21.8%
Gender	Male	208	55.3%
	Female	168	44.7%
Educational level	Secondary school and below	117	31.1%
	Associate degree	114	30.3%
	Bachelor degree	107	28.5%
	Master degree	38	10.1%
Occupation	Rural entrepreneur	84	22.3%
	B&B operator	48	12.8%
	Service worker	76	20.2%
	Farmer	62	16.5%
	Other	106	28.2%
Length of residence	<1 year	61	16.2%
	1–3 years	135	35.9%
	3–5 years	106	28.2%
	>5 years	74	19.7%

.

4. Results

Based on the results of in-depth interview theme clustering and combining the Hierarchy of Effects Model, this section constructs a structural equation model of the influence of rural community co-build willingness and conducts empirical research based on questionnaire data. At the same time, in order to clarify the differences in the driving mechanisms of co-build willingness between different groups of immigrants and local villagers, this study sets the migrant group questionnaire data as Group A and the local villagers as Group B to conduct group experiments.

4.1. The Measurement Model Assessment

The study employed IBM SPSS Statistics 27.0 (IBM Corp., Armonk, NY, USA) for confirmatory factor analysis and descriptive statistics, and further applied IBM SPSS Amos 28.0 (IBM Corp., Armonk, NY, USA) for SEM analysis. As shown in the model fit test results in

Table 4. The values of fit indices.

Fit Indices	Standard	Group A Results	Group B Results
CMIN/DF	1–3 is excellent, 3–5 is good	1.651	1.735
RMSEA	<0.05 is excellent, <0.08 is good	0.045	0.06
NFI	>0.9 is excellent, >0.8 is good	0.934	0.9
IFI	>0.9 is excellent, >0.8 is good	0.973	0.955
CFI	>0.9 is excellent, >0.8 is good	0.973	0.955

, first, the CMIN/DF values for Model A were 1.651 and 1.735 for Group A and Group B, respectively, both within the excellent range of 1–3, indicating that the model fit for both groups was excellent. Second, the RMSEA value for Group A was 0.045, and for Group B, it was 0.06, both below 0.05, falling within the excellent range. Additionally, the normalized fit index (NFI) for Group A was 0.934, and for Group B, it was 0.9, although Group B was slightly below 0.9, both groups’ NFI values indicated good model fit; The incremental fit index (IFI) for Group A was 0.973, and for Group B it was 0.955, both above 0.9; the comparative fit index (CFI) for Group A was 0.973, and for Group B it was 0.955, also above 0.9, further confirming that the model fit for both groups of data is excellent.

The KMO test is used to assess the suitability of data for factor analysis, with values ranging from 0 to 1. As shown in

Table 5. KMO and Bartlett’s inspection.

Level 1 Indicators	Level 2 Indicators	Group A Results	Group B Results
KMO		0.820	0.882
Bartlett’s sphericity	spherical test	1994.331	2898.109
	df-value	210	210
	p-value	<0.001	0.000

, the KMO value for Group A is 0.820, while the KMO value for Group B is 0.882. Both values are above 0.8, indicating that both sets of data have good sampling adequacy and are suitable for factor analysis. Bartlett’s sphericity test is a standard procedure used to verify the suitability of data for factor analysis by examining the correlation matrix. It measures whether variables are independent of each other. If the significance level of the test is less than 0.05, it indicates that there is a correlation between variables, making them suitable for factor analysis. The approximate chi-square value for Group A is 1994.331, with 210 degrees of freedom and a significance level less than 0.001. For Group B, the approximate chi-square value is 2898.109, with the same 210 degrees of freedom and a significance level of 0.000. These results indicate that there is a significant correlation between the variables in the two groups of data, further supporting the applicability of factor analysis. The results confirm the reliability of the data in this study and validate the effectiveness of the research findings.

4.2. Confirmatory Factor Analysis

When conducting SEM analysis, confirmatory factor analysis (CFA) is an important statistical method used to verify whether the variable structure proposed by researchers matches the actual data. Cronbach’s alpha is a measure of internal consistency for a scale, with values ranging from 0 to 1. Typically, a value above 0.7 indicates acceptable internal consistency, while a value above 0.8 indicates good internal consistency. The analysis results are shown in

Table 6. Results of confirmatory factor analysis.

Variable Dimension	Measurement Items	Group A α	Group B α
Spatial–environmental continuity	Diversified environmental features	0.870	0.852
	Composite spatial functions
	Social service levels
Cultural landscape transition	Cultural symbol translation	0.926	0.933
	Consumption scenario iteration
	Dynamic Revitalization of Traditional Festivals
Social interaction embeddedness	Interaction between cross-groups	0.889	0.866
	Immigrant Community Development
	Neighbor Relations Adjustment
New rural livability	Inclusiveness of consumer services	0.852	0.890
	Optimization of residential facilities
	Rural Innovation and Employment Attraction
Community identity	Cultural subjectivity perception	0.791	0.863
	Consensus on community values
	Collective dignity maintenance tendency
Sense of belonging	Emotional anchoring depth	0.827	0.821
	Perception of community acceptance
	Long-term attachment
Community Co-build willingness	Tendency to environmental creation	0.819	0.849
	Cultural capital investment
	Public affairs involvement

. The reliability coefficients of spatial–environmental continuity, cultural landscape transition, social interaction embeddedness, new rural livability, community identity, sense of belonging, community co-build willingness, and various secondary indicators are all between 0.7 and 1. This indicates that the scales used in this study have high internal consistency and good reliability.

In this study, construct consistency (CR) was used to measure the internal consistency of latent variables, average variance extracted (AVE) was used to measure the proportion of variance in latent variables explained by their indicator variables, and standard regression coefficients reflected the strength of the relationship between observed variables and latent variables. As shown in

Table 7. Model Convergent Validity.

Path Relationship	Group A Standard Regression Coefficient	Group A AVE	Group A CR	Group B Standard Regression Coefficient	Group B AVE	Group B CR
SEC1 <--- SEC	0.817	0.682	0.865	0.786	0.658	0.852
SEC2 <--- SEC	0.836			0.817
SEC3 <--- SEC	0.824			0.829
CLT1 <--- CLT	0.867	0.818	0.931	0.859	0.824	0.934
CLT2 <--- CLT	0.926			0.941
CLT3 <--- CLT	0.919			0.922
CIE1 <--- CIE	0.890	0.754	0.902	0.822	0.685	0.867
CIE2 <--- CIE	0.867			0.768
CIE3 <--- CIE	0.847			0.888
NRL1 <--- NRL	0.805	0.704	0.877	0.819	0.731	0.890
NRL2 <--- NRL	0.848			0.871
NRL3 <--- NRL	0.863			0.873
CI1 <--- CI	0.832	0.631	0.837	0.862	0.673	0.861
CI2 <--- CI	0.768			0.770
CI3 <--- CI	0.781			0.827
SOB1 <--- SOB	0.789	0.617	0.828	0.768	0.597	0.816
SOB2 <--- SOB	0.802			0.795
SOB3 <--- SOB	0.765			0.754
CW1 <--- CW	0.867	0.633	0.837	0.840	0.653	0.850
CW2 <--- CW	0.767			0.784
CW3 <--- CW	0.747			0.800

, in Group A, all latent variables had AVE values exceeding the threshold of 0.5, indicating that the model has good convergent validity. The AVE values in Group B also showed a similar trend, with all latent variables having AVE values greater than 0.5. According to the data recorded in this study, in Group A, all CR values of the latent variables are greater than 0.8, indicating high reliability. The CR values in Group B also show the same trend. According to the reference standards, CR values range from 0 to 1, with higher values indicating greater reliability. Based on these results, the convergent validity of the model is excellent in both Group A and Group B, indicating that the indicator variables in the model effectively explain their corresponding latent variables.

4.3. Moderating Effect Test

In the path hypothesis testing conducted in this study, as shown in

Table 8. Group A path analysis results.

Implicit Variable	Path	Independent Variable	Estimate	S.E. (Standard Error)	C.R. (Critical Ratio)
CI	<---	SEC	0.139	0.065	2.144
CI	<---	CLT	0.276	0.054	5.066
CI	<---	CIE	0.119	0.056	2.142
CI	<---	NRL	0.165	0.053	3.132
SOB	<---	SEC	0.161	0.062	2.581
SOB	<---	CLT	0.168	0.052	3.241
SOB	<---	CIE	0.209	0.054	3.846
SOB	<---	NRL	0.228	0.051	4.426
CW	<---	CI	0.168	0.07	2.404
CW	<---	SOB	0.442	0.075	5.859

and

Table 9. Group B path analysis results.

Latent Variable	Path	Independent Variable	Estimate	S.E. (Standard Error)	C.R. (Critical Ratio)
CI	<---	SEC	0.259	0.086	3.031
CI	<---	CLT	0.262	0.071	3.701
CI	<---	CIE	0.16	0.071	2.252
CI	<---	NRL	0.066	0.069	0.956
SOB	<---	SEC	0.09	0.079	1.144
SOB	<---	CLT	0.168	0.066	2.536
SOB	<---	CIE	0.288	0.069	4.148
SOB	<---	NRL	0.201	0.066	3.023
CW	<---	CI	0.294	0.067	4.422
CW	<---	SOB	0.609	0.083	7.314

and Figure 4 the standardized regression coefficients (β) represent the direct effect values between variables. These coefficients are also referred to as path coefficients. Using a significance level of 0.05, researchers can determine whether a direct effect exists between two variables. If a significant relationship exists, the standardized regression coefficient indicates the strength of this effect.

In the data of Group A shown in Table 8 and Figure 4, spatial–environmental continuity had a significant positive effect on community identity (Estimate = 0.139, p = 0.032, β = 0.142) and sense of belonging (Estimate = 0.161, p = 0.01, β = 0.169). Cultural landscape transition had a significant positive effect on community identity (Estimate = 0.276, p < 0.001, β = 0.358) and sense of belonging (Estimate = 0.168, p = 0.001, β = 0.222). Social interaction embeddedness had a significant positive effect on community identity (Estimate = 0.119, p = 0.032, β = 0.14) and sense of belonging (Estimate = 0.209, p < 0.001, β = 0.25). New rural livability has a significant positive effect on community identity (Estimate = 0.165, p = 0.002, β = 0.188) and sense of belonging (Estimate = 0.228, p < 0.001, β = 0.266). Community identity has a significant positive effect on community co-build willingness (Estimate = 0.168, p = 0.016, β = 0.163), and sense of belonging has a significant positive effect on community co-build willingness (Estimate = 0.442, p < 0.001, β = 0.419).

In the data of Group B shown in Table 9 and Figure 4, spatial–environmental continuity had a significant positive effect on community identity (Estimate = 0.259, p = 0.002, β = 0.248), but no significant effect on sense of belonging (Estimate = 0.09, p = 0.253). Cultural landscape transition had a significant positive effect on community identity (Estimate = 0.262, p < 0.001, β = 0.333) and sense of belonging (Estimate = 0.168, p = 0.011, β = 0.227). Social interaction embeddedness had a significant positive effect on community identity (Estimate = 0.16, p = 0.024, β = 0.18) and sense of belonging (Estimate = 0.288, p < 0.001, β = 0.343). New rural livability had no significant effect on community identity (Estimate = 0.066, p = 0.339), but had a significant positive effect on sense of belonging (Estimate = 0.201, p = 0.003, β = 0.242). Community identity had a significant positive effect on the community co-build willingness (Estimate = 0.294, p < 0.001, β = 0.312), and the sense of belonging had a significant positive effect on the community co-build willingness (Estimate = 0.609, p < 0.001, β = 0.609).

4.4. Construction of the Artificial Neural Network (ANN) Model

In order to further optimize the model and improve its predictive accuracy, this section combines the SEM analysis results with the artificial neural network (ANN) method to construct a SEM-ANN-based model of the willingness of rural immigrants and local villagers to co-build communities. Referring to the study by [75], four artificial neural network models A, B, C, and D were constructed based on SEM results and ANN principles. Models A and B were constructed using immigrant data, while models C and D were constructed using local villager data. As shown in Figure 5, the input layer of model A includes spatial–environmental continuity (SEC), cultural landscape transition (CLT), social interaction embeddedness (CIE), and new rural livability (NRL), and the output layer includes community identity (CI) and sense of belonging (SOB). The input layer of model B consists of community identity (CI) and sense of belonging (SOB), and the output layer is community co-build willingness (CW). The input layer of Model C includes spatial environmental continuity (SEC), cultural landscape transition (CLT), social interaction embeddedness (CIE), and new rural livability (NRL), and the output layer is community identity (CI) and sense of belonging (SOB). The input layer of Model D includes community identity (CI) and sense of belonging (SOB), and the output layer is community co-build willingness (CW).

4.5. Root Means Square Error Validation

To systematically evaluate the predictive accuracy of artificial neural network (ANN) models, this study employed tenfold cross-validation to assess the performance of models A, B, C, and D. This method first divides the entire dataset into ten mutually exclusive subsets. Through iterative sampling, a training-testing combination is constructed: each time, nine subsets (90% of the samples) are randomly selected for model training, and the remaining one subset (10% of the samples) is used as the test set. This process is repeated ten times to ensure that all subsets are used as test data once, thereby comprehensively covering the data distribution characteristics. Based on this validation framework, the root mean square error (RMSE) metric is quantitatively calculated for each ANN model in the prediction task. As shown in

Table 10. Root means square error test.

	Model A		Model B		Model C		Model D
	Input: SEC, CLT, CIE, NRL		Input: CI, SOB		Input: SEC, CLT, CIE, NRL		Input: CI, SOB
	Output: CI, SOB		Output: CW		Output: CI, SOB		Output: CW
Neural network	Training	Testing	Training	Testing	Training	Testing	Training	Testing
ANN1	0.169	0.179	0.157	0.181	0.184	0.163	0.104	0.105
ANN2	0.177	0.207	0.165	0.117	0.190	0.180	0.116	0.093
ANN3	0.171	0.159	0.158	0.173	0.178	0.210	0.116	0.092
Mean	0.172	0.182	0.160	0.157	0.184	0.184	0.112	0.097
SD (Standard Deviation)	0.004	0.024	0.004	0.035	0.006	0.024	0.007	0.007

, the RMSE values range from 0.092 to 0.210, indicating that the prediction errors of all models are relatively low. These results demonstrate that artificial neural network models exhibit strong predictive capabilities for the test data, providing reliable and accurate support for subsequent analysis and decision-making processes.

4.6. Sensitivity Analysis

Through sensitivity analysis using artificial neural networks (ANN), this study identified differences in the relative importance of covariates on the target variable across different models. As shown in

Table 11. Analysis of the importance of normalization in ANN models of immigrants.

	Model A				Model B
	Output: CI, SOB				Output: CW
Neural network	SEC	CLT	CIE	NRL	CI	SOB
ANN1	0.142	0.339	0.100	0.418	0.065	0.935
ANN2	0.262	0.379	0.025	0.333	0.489	0.511
ANN3	0.142	0.383	0.119	0.356	0.249	0.761
Average relative importance	0.182	0.367	0.081	0.369	0.268	0.736
Normalized relative importance (%)	49.323	99.455	21.976	100.000	36.413	100.000

. In Model A, the importance rankings of the predictor variables are NRL (100.000%), CLT (99.455%), SEC (49.323%), and CIE (21.976%). In Model B, the importance of SOB (100.000%) far exceeds that of CI (36.413%). As shown in

Table 12. Analysis of the importance of normalization in ANN models of local villager.

	Model C				Model D
	Output: CI, SOB				Output: CW
Neural network	SEC	CLT	CIE	NRL	CI	SOB
ANN1	0.203	0.273	0.347	0.176	0.420	0.580
ANN2	0.175	0.286	0.363	0.176	0.436	0.564
ANN3	0.185	0.302	0.315	0.197	0.531	0.469
Average relative importance	0.188	0.287	0.342	0.183	0.462	0.538
Normalized relative importance (%)	54.971	83.918	100.000	53.509	85.859	100.000

. Model C analysis shows that the importance ranking is CIE (100.000%), CLT (83.918%), SEC (54.971%), and NRL (53.509%). Model D shows that the variable importance is SOB (100.000%), CI (85.859%). These differences reveal the sensitivity changes in variable importance across different models and help this study understand which variables have a more significant impact on the dependent variables in each model.

5. Discussion

The data results show that the SEM and ANN models are relatively consistent in the ranking of the importance of influencing factors. In the SEM-based model of the willingness of immigrants and local villagers to co-build communities, except for H1b and H4a in Group B, all hypotheses were successfully tested. Among them, the variables that had the greatest impact on the willingness of immigrants to co-build communities were cultural landscape transition, new rural livability, and sense of belonging. The variables that had the greatest impact on the willingness of local villagers to co-build communities were cultural landscape transition, social interaction embeddedness, and sense of belonging. It is noteworthy that the few paths in the model that failed to pass the significance test do not negate the importance of the variables themselves. Rather, they reveal fundamental differences between the two groups in terms of landscape perception and behavioral motivation, providing crucial clues for subsequent mechanism analysis. Firstly, in terms of SEC’s influence on SOB, a sense of belonging is rooted more in long-term accumulated family ties and neighborhood networks than in a single landscape continuity. The SEC continues to reinforce their perception of the village as a community and their sense of identity, yet it struggles to exert any significant additional influence beyond their already high and stable sense of belonging. Secondly, with regard to the impact of the NRL on CI, this initiative primarily enhances local villagers’ practical reliance on their community by improving living conditions and livelihood opportunities. Its principal pathway of influence is: New Rural Livability—Sense of Belonging—Co-build willingness. In contrast, local residents’ assessment of community identity relies more heavily on the continuity of cultural symbols and the embedding of social relationships.

5.1. Verification of the Impact Mechanism of Gentrified Landscapes

The study proved that the gentrified landscape of rural areas significantly and positively influences the community identity and sense of belonging of immigrants and local villagers by integrating the four dimensions of physical landscape, symbolic landscape, social landscape, and living landscape, thereby promoting the cultivation of their willingness to participate in community co-building. For migrant groups (Group A), the formation of community identity is centered on the cultural landscape transition, followed by the new rural livability. This finding is supported by Keleg et al., who found that there is a relationship between landscape characteristics, aesthetic experiences, and the socio-cultural identity of a community [76]. This shows that immigrants, through gentrified landscape projects, modernize traditional cultural symbols, preserving their cultural genes while giving them market value. This strategy of ‘cultural capital transformation’ [77] enables immigrants to establish a sense of community identity through innovation. For the local villagers (Group B), community identity depends on the degree of cultural landscape transition and spatial–environmental continuity. Qian et al. pointed out that cultural capital rooted in the minds and practices of local people and the residential environment help to form new local meanings that support community identity [78]. Local villagers have a strong emotional attachment to physical spaces (such as ancestral halls and farmland), and their identity is rooted in the cultural capital of traditional symbols. Therefore, gentrified landscapes need to balance protection and development through functional zoning, leaving corresponding space for the coexistence of both in terms of symbolic meaning and spatial planning to avoid triggering cultural defensive reactions. The driving mechanism of belonging differs from the logical construction of community identity. For immigrant groups (Group A), a sense of belonging is mainly driven by new rural livability and social interaction embeddedness. This finding suggests that when gentrified landscapes become the foundation for immigrants’ income growth and social capital accumulation in a place, they will have a stronger sense of belonging. Greinke and Rammelmeier point out that if individuals in a community have particularly strong roots in one place or stronger roots than in another, they are more likely to attempt to maintain community participation there rather than in a place where they lack roots or a sense of belonging [79]. The sense of belonging among local villagers stems more from social interaction embeddedness, a finding supported by Castle and Grant, who point out that social connections are important for fostering a sense of belonging [80].

5.2. Differentiated Approaches to the Rural Co-Build Willingness

The study found significant differences in the driving factors behind the willingness of immigrants and local villagers to co-build communities. The willingness of migrant groups to co-build communities follows the principle of benefit-driven participation, while the participation logic of local villagers is rooted in cultural continuity. This dual track reveals that the community participation of local villagers is an extension of cultural inertia, while the co-building behavior of immigrants is essentially a strategic practice of viewing community co-building as a means of obtaining development resources. The core driving path of the community co-build willingness of the migrant group (Group A) is: new rural livability—sense of belonging—community co-build willingness. The core driving path of the community co-build willingness of local villagers (Group B) is: social interaction embeddedness—sense of belonging—community co-build willingness. Both reflect the importance of a sense of belonging in the community co-build willingness mechanism, which is basically consistent with existing academic research results [51,81,82]. However, the most important drivers of a sense of belonging among immigrants and local villagers are the new rural livability and social interaction embeddedness, respectively, which may explain the structural differences in the levels of needs and resource dependence of immigrants and local villagers in community participation. For immigrants, the dominant role of new rural livability reflects their priority focus on material security and development opportunities when migrating or integrating into a new community. The gentrified rural landscape provides immigrants with entry points for economic participation through the integration of culture and tourism and the creation of employment opportunities. This reflects the key role of economic embeddedness in the context of population mobility [83]. The sensitivity of local villagers to social interaction embeddedness may be rooted in the social capital accumulated through their long-term community life. Social networks, traditional norms, and mutual aid relationships constitute the core source of their sense of belonging. The intervention of gentrified landscapes does not necessarily sever the traditional social networks of local villagers, but may instead strengthen the role of social interaction embeddedness through organized operations. For example, traditional handicraft workshops and festival activities transform local villagers’ local knowledge into social capital, preserving their dominance in cultural capital [77]. This process protects the continuity of local culture while reinforcing trust within the community.

5.3. Policy Suggestions

Based on the research and analysis in this study, the following recommendations are proposed to stimulate immigrants and local villagers’ community identity and sense of belonging to the gentrified landscape, generate co-build willingness in the community, and promote the sustainable development of rural communities.

First, in order to unleash the integration potential of gentrified landscapes, planning practices need to focus on the cultural landscape transition of gentrified landscapes. Through the innovative revitalization of traditional elements, modern functions can be injected while continuing the authentic narrative, thereby simultaneously satisfying the cultural innovation demands of immigrants and the heritage protection needs of villagers, effectively enhancing the “sense of belonging” of rural communities, supplementing the collective memory connections between different groups, and thereby strengthening the community identity of rural community.

Second, policy design needs to anchor the central role of emotional transformation in the sense of belonging to the countryside. Sense of belonging plays a central role in community integration. Therefore, the creation of a sense of belonging should be achieved through two-pronged intervention: on the one hand, protecting material carriers of collective memory such as ancestral halls and ancient trees to maintain the historical legibility of the spatial environment; on the other hand, designing cross-group production and collaboration scenarios to deeply embed social interaction in daily life. When immigrants and local villagers participate in collaborative practices in a continuous field, the intensity of their emotional attachment and community co-build willingness will increase.

Finally, in response to the differing motivational logic of immigrants and local villagers, a dual-track empowerment strategy should be implemented: immigrants should focus on enhancing their income-generating capacity in new industries, with a focus on establishing transparent property rights transfer systems and entrepreneurship support policies to strengthen the sustainable profitability of their operational landscapes, thereby effectively converting economic benefits into community belonging. Local villagers should focus on preserving spatial memory and social culture. The priority is to establish a graded protection system for historical carriers and institutionalize safeguards for cultural rituals. This cultural continuity, catalyzed by the coexistence of spatial continuity and bodily practices, ultimately triggers co-construction actions driven by the self-awareness of “guardians”, thereby effectively enhancing the integration of rural communities.

5.4. Limitations of the Study and Future Research Needs

This study has made certain theoretical contributions to the study of rural community building and provided practical guidance for the construction of gentrified landscapes in rural areas to meet the needs of more groups and enhance the cultural influence of rural areas. However, there are still some limitations. First, the data in this study is time-sensitive, and it will be necessary to track the impact of gentrified landscapes on cultural identity and community identity, especially the identity changes in young villagers, in the long term. Second, the construction of rural communities discussed in this study is a broad concept that covers diverse subjects such as local villagers, immigrants, returning entrepreneurs, and migrant workers. Due to their different needs and experiences, the incentives for different subjects to participate in community co-building are also worth exploring in depth. The theme clustering results obtained in this study should be combined with other theories to fully understand the weight of factors under different theories and their interactions. In short, future research needs to consider these factors to further improve the results of this study.

6. Conclusions

Contemporary rural areas face dual challenges of preserving local identity while fostering innovative development. By examining residents’ perceptions, emotional attachment, and co-build willingness within gentrified rural landscapes, this study provides valuable empirical evidence and planning insights for rural regeneration and community integration. Firstly, based on field research, this study applied LDA modeling to identify key perception themes influencing community co-build willingness. Subsequently, using Shecun Village in Nanjing as a case study, SEM combined with ANN analysis was employed to robustly examine how Spatial–environmental continuity, Cultural landscape transition, social interaction embeddedness, and New rural livability jointly shape community identity Sense of belonging and co-build willingness. Based on these findings, this study proposes a series of practical strategic recommendations aimed at enhancing the willingness of local villagers and immigrants to engage in community co-creation, thereby increasing the village’s overall appeal and long-term development potential.

This study offers new perspectives and methods for understanding and promoting sustainable integration in gentrified rural communities. It emphasizes the need to balance cultural continuity and social cohesion in rural landscape renewal. Using an integrated LDA–SEM–ANN framework, the study validates the key roles of landscape perception and emotional attachment in shaping co-build willingness. The findings provide actionable strategies for rural governance and community development. The proposed framework is applicable not only to China’s rural regeneration but also to global rural revitalization efforts.