Search Results (11)

Manual code review is essential for software quality, but often slows down development cycles due to the high time demands on developers. In this study, we propose an automated solution for Python (version 3.13) projects that generates code review comments by combining Large Language Models (LLMs) with Retrieval-Augmented Generation (RAG). To achieve this, we first curated a dataset from GitHub pull requests (PRs) using the GitHub REST Application Programming Interface (API) (version 2022-11-28) and classified comments into semantic categories using a semi-supervised Support Vector Machine (SVM) model. During the review process, our system uses a vector database to retrieve the top-k most relevant historical comments, providing context for a diverse spectrum of open-weights LLMs, including DeepSeek-Coder-33B, Qwen2.5-Coder-32B, Codestral-22B, CodeLlama-13B, Mistral-Instruct-7B, and Phi-3-Mini. We evaluated the system using a multi-step validation that combined standard metrics (BLEU-4, ROUGE-L, cosine similarity) with an LLM-as-a-Judge approach, and verified the results through targeted human review to ensure consistency with expert standards. The findings show that retrieval augmentation improves feedback relevance for larger models, with DeepSeek-Coder’s alignment score increasing by 17.9% at a retrieval depth of k = 3. In contrast, smaller models such as Phi-3-Mini suffered from context collapse, where too much context reduced accuracy. To manage this trade-off, we built a hybrid expert system that routes each task to the most suitable model. Our results indicate that the proposed approach improved performance by 13.2% compared to the zero-shot baseline (k = 0). In addition, our proposed system reduces hallucinations and generates comments that closely align with the standards expected from the experts. Full article

Crowdsourcing Software Development (CSD) platforms, i.e., TopCoder, function as intermediaries connecting clients with developers. Despite employing systematic methodologies, these platforms frequently encounter high task abandonment rates, with approximately 19% of projects failing to meet satisfactory outcomes. Although existing research has focused on task scheduling, developer recommendations, and reward mechanisms, there has been insufficient attention to the support of platform moderators, or copilots, who are essential to project success. A critical responsibility of copilots is estimating project duration; however, manual predictions often lead to inconsistencies and delays. This paper introduces an innovative machine learning approach designed to automate the prediction of project duration on CSD platforms. Utilizing historical data from TopCoder, the proposed method extracts pertinent project attributes and preprocesses textual data through Natural Language Processing (NLP). Bidirectional Encoder Representations from Transformers (BERT) are employed to convert textual information into vectors, which are then analyzed using various machine learning algorithms. Zero-shot learning algorithms exhibit superior performance, with an average accuracy of 92.76%, precision of 92.76%, recall of 99.33%, and an f-measure of 95.93%. The implementation of the proposed automated duration prediction model is crucial for enhancing the success rate of crowdsourcing projects, optimizing resource allocation, managing budgets effectively, and improving stakeholder satisfaction. Full article

Competitive Crowdsourcing Software Development (CCSD) is popular among academics and industries because of its cost-effectiveness, reliability, and quality. However, CCSD is in its early stages and does not resolve major issues, including having a low solution submission rate and high project failure risk. Software development wastes stakeholders’ time and effort as they cannot find a suitable solution in a highly dynamic and competitive marketplace. It is, therefore, crucial to automatically predict the success of an upcoming software project before crowdsourcing it. This will save stakeholders’ and co-pilots’ time and effort. To this end, this paper proposes a well-known deep learning model called Bidirectional Encoder Representations from Transformers (BERT) for the success prediction of Crowdsourced Software Projects (CSPs). The proposed model is trained and tested using the history data of CSPs collected from TopCoder using its REST API. The outcomes of hold-out validation indicate a notable enhancement in the proposed approach compared to existing methods, with increases of 13.46%, 8.83%, and 11.13% in precision, recall, and F1 score, respectively. Full article

Room usage semantics in models of large indoor environments such as public buildings and business complex are critical in many practical applications, such as health and safety regulations, compliance, and emergency response. Existing models such as IndoorGML have very limited semantic information at room level, and it remains difficult to capture semantic knowledge of rooms in an efficient way. In this paper, we formulate the task of generating rooms usage semantics as a special case of room classification problems. Although methods for room classification tasks have been developed in the field of social robotics studies and indoor maps, they do not deal with room usage and occupancy aspects of semantics, and they ignore the value of furniture objects in understanding room usage. We propose a method for generating room usage semantics based on the spatial configuration of room objects (e.g., furniture, walls, windows, doors). This method uses deep learning architecture to support a room usage classifier that can learn spatial configuration features directly from semantically labelled point cloud (SLPC) data that represent room scenes with furniture objects in place. We experimentally assessed the capacity of our method in classifying rooms in office buildings using the Stanford 3D (S3DIS) dataset. The results showed that our method was able to achieve an overall accuracy of 91% on top-level room categories (e.g., offices, conference rooms, lounges, storage) and above 97% accuracy in recognizing offices and conference rooms. We further show that our classifier can distinguish fine-grained categories of of offices and conference rooms such as shared offices, single-occupancy offices, large conference rooms, and small conference rooms, with comparable intelligence to human coders. In general, our method performs better on rooms with a richer variety of objects than on rooms with few or no furniture objects. Full article

Videos are arguably the most important and frequently used instructional resource in massive open online courses (MOOCs). Recent research has explored learners’ perceptions and preferences regarding MOOC instructional videos. However, these studies are often limited to a small number of specific courses, and few grounded theory studies have been undertaken to investigate this topic. In the present study, a multiple-coder research methodology was adopted to analyze 4534 learner reviews of MOOCs in 14 categories. The study aimed to identify key characteristics associated with learners’ favorable perceptions of MOOC videos, types of supplemental or in-video resources learners perceive helpful to support MOOC video use, and video production features learners value. Results revealed that (a) “organized”, “detailed”, “comprehensible”, “interesting”, and “practical” were the top five important characteristics associated with learners’ favorable perceptions of MOOC videos; (b) learners perceived “presentation slides”, “reading materials”, “post-video assessments”, “embedded questions”, and “case studies” as helpful resources to support their utilization of MOOC videos; and (c) learners found “duration” a more salient production feature than “editing”, “resolution”, “subtitles”, “music”, or “voice”. The findings present implications for MOOC video design and foundations for future research avenues. Full article

The experience linked to the COVID-19 emergency constituted a turning point in the biography of most Italians. The suspension of usual activities, the redefinition of life contexts and the restriction of relationships have opened up wide spaces and time for thinking and reflecting on oneself, which may have triggered processes of redefinition of personal identity. The general aim of this study was to explore the impact of pandemic on daily life in the life span, in order to support the hypothesis that the pandemic experience could be considered a disequilibrating life-event and a turning point in the biography of most Italians. A mixed research approach was adopted, with 14 closed and open questions created ad hoc. 41 participants (87% women, average age 40.71), resident in the Campania region, in southern Italy, responded to the online written interview. The data were analyzed by two independent coders, using categorical content analysis with a top-down approach. Membership of the different age groups (young adults, adults, elderly) was assessed as a comparison variable. Findings qualify pandemic-related experiences as a disequilibrating life event, potentially capable of activating, alongside emotionally dense experiences, adaptive and functional resources for identity reconsideration, with differences being age based. The dimensions of change, the affective dimensions, the resources and the areas of risk identified, allowed us to identify three different clusters, showing a differentiation according to age groups, which identifies young adults and the elderly as the subjects most at risk. Full article

Increasing evidence has suggested that microRNAs (miRNAs) are significant in research on human diseases. Predicting possible associations between miRNAs and diseases would provide new perspectives on disease diagnosis, pathogenesis, and gene therapy. However, considering the intrinsic time-consuming and expensive cost of traditional Vitro studies, there is an urgent need for a computational approach that would allow researchers to identify potential associations between miRNAs and diseases for further research. In this paper, we presented a novel computational method called SMMDA to predict potential miRNA-disease associations. In particular, SMMDA first utilized a new disease representation method (MeSHHeading2vec) based on the network embedding algorithm and then fused it with Gaussian interaction profile kernel similarity information of miRNAs and diseases, disease semantic similarity, and miRNA functional similarity. Secondly, SMMDA utilized a deep auto-coder network to transform the original features further to achieve a better feature representation. Finally, the ensemble learning model, XGBoost, was used as the underlying training and prediction method for SMMDA. In the results, SMMDA acquired a mean accuracy of 86.68% with a standard deviation of 0.42% and a mean AUC of 94.07% with a standard deviation of 0.23%, outperforming many previous works. Moreover, we also compared the predictive ability of SMMDA with different classifiers and different feature descriptors. In the case studies of three common Human diseases, the top 50 candidate miRNAs have 47 (esophageal neoplasms), 48 (breast neoplasms), and 48 (colon neoplasms) are successfully verified by two other databases. The experimental results proved that SMMDA has a reliable prediction ability in predicting potential miRNA-disease associations. Therefore, it is anticipated that SMMDA could be an effective tool for biomedical researchers. Full article

Circular business model innovation offers a path for the transformation of companies, enhancing resource productivity and efficiency, while also contributing to sustainable development. These fundamental changes in business are accompanied by a variety of challenges and barriers. To support companies on their journey, only a few studies have investigated the critical success factors for circular business model innovation through literature analysis. To contribute to this research, in this study, a methodological approach, mainly based on expert interviews, is proposed to gain in-depth insight into critical success factors for circular business model innovation. As a result, a framework covering critical success factors for circular business model innovation is developed, comprising nine top-codes and 37 sub-codes, and an analysis of each factor’s contribution to the UN’s Sustainable Development Goals is performed. The study thereby extends the theoretical basis for further research on circular business model innovation, as well as identifies their practical implications. Full article

Viewing online pornography is common among US adults, with mixed-sex threesome (MST) videos being one of the top 10 most popular categories of pornography for both men and women. The current content analysis applied sexual script theory to understand the themes present in these mixed-sex threesome videos. Independent coders viewed a total of 50 videos (25 MMF and 25 FFM) at each timepoint (2012, 2015, 2020) and coded for different sexual behaviors and themes in each video. By examining both same-sex (female–female, male–male) and other-sex (female–male) behaviors, as well as themes of aggression and sexual initiation in different videos and across three timepoints, it was determined that other-sex behaviors are more common in MST videos than same-sex behaviors. Same-sex behaviors between two female actors were more common than same-sex behaviors between two male actors. Aggression was a common theme in videos, with male actors being more aggressive on average than female actors. Most of these trends did not change across 8 years, suggesting that the impacts of traditional sexual scripts are pervasive in pornography, even in current online content. Important implications for both researchers and clinical professionals are discussed. Full article

Disordered social media use, often referred to as “social media addiction”, has not been officially recognized by medical bodies such as the American Psychiatric Association or the World Health Organization. However, websites still present information to laypeople on how to treat and manage social media addiction, which can pose the risk of spreading low quality or incorrect information. As such, we aimed to assess how the most popular social media addiction websites present information across multiple metrics. We conducted an in-depth online search to identify the top social media addiction websites in November 2019 (N = 23). Websites were separated into four distinct classifications: (1) treatment/therapy/medical; (2) informational; (3) news article; and (4) blog/essay. Based on previous website analysis research, three trained coders evaluated these websites on six metrics: (1) design; (2) credibility; (3) accessibility; (4) literacy; (5) engagement; and (6) social media addiction content. Design features were the top-rated metric across all websites, followed by credibility. Websites scored the lowest for the engagement and social media addiction content metrics. Across website classifications, scores for social media addiction content varied greatly, with blog/essay websites ranking the lowest and informational websites ranking the highest. Our findings provide necessary information for both patients and healthcare providers, apprising these individuals and the field about the current online health information landscape for disordered social media use. Full article

Crowdsourcing has the advantages of being cost-effective and saving time, which is a typical embodiment of collective wisdom and community workers’ collaborative development. However, this development paradigm of software crowdsourcing has not been used widely. A very important reason is that requesters have limited knowledge about crowd workers’ professional skills and qualities. Another reason is that the crowd workers in the competition cannot get the appropriate reward, which affects their motivation. To solve this problem, this paper proposes a method of maximizing reward based on the crowdsourcing ability of workers, they can choose tasks according to their own abilities to obtain appropriate bonuses. Our method includes two steps: Firstly, it puts forward a method to evaluate the crowd workers’ ability, then it analyzes the intensity of competition for tasks at Topcoder.com—an open community crowdsourcing platform—on the basis of the workers’ crowdsourcing ability; secondly, it follows dynamic programming ideas and builds game models under complete information in different cases, offering a strategy of reward maximization for workers by solving a mixed-strategy Nash equilibrium. This paper employs crowdsourcing data from Topcoder.com to carry out experiments. The experimental results show that the distribution of workers’ crowdsourcing ability is uneven, and to some extent it can show the activity degree of crowdsourcing tasks. Meanwhile, according to the strategy of reward maximization, a crowd worker can get the theoretically maximum reward. Full article