An Improved Case-Based Reasoning Model for Simulating Urban Growth

Round 1

Reviewer 1 Report

The subject of the article is interesting. After reading it carefully, the following points were formulated. 

- What is new to knowledge from the problem presented in the paper? 

- Table 1: Indexes listed in col.1 require more explanation because the description in col.2 "Process mode" is too general; 

- It is recommended to edit the text so that the references to Tables and Figures are closer to these objects; 

- What is the advantage of the model presented in the paper over other models? 

Author Response

Response to Reviewer 1 Comments

Thanks for the reviewer's comments. We have responded to the comments in the following.

Point 1: What is new to knowledge from the problem presented in the paper?


Response 1: Many scholars have used a variety of prediction methods to model land expansion in different cities, including cellular automata (CA) [1, 2], the Flus model [3], the multi-agent system[4, 5], the Clue-S model [6, 7], and the system dynamics model[8]. However, the complex urban pattern evolution mechanism makes it difficult to extract relevant knowledge. This makes these models difficult to interpret and reduces their reliability. Therefore, it is necessary to adopt new research methods to break through the existing bottlenecks.

CBR can even achieve quantitative analysis and prediction without needing to scrutinize the mechanism, so CBR can effectively break through the bottleneck of knowledge acquisition present in traditional models.

Few studies have examined the spatial patterns or evolution processes of geographical phenomena using CBR. Among them, only one has proposed a CBR model for simulating land use changes [9]. However, this model struggles to directly describe how the process of urban growth leads to observed changes. Furthermore, the single retrieval result leads to a low level of reliability regarding inferences, and the lack of a time control factor in the fuzzy inference cycle makes it difficult to determine the simulation occurrence when simulating urban growth. The present study therefore improves the CBR model regarding the simulation of land use changes and constructs an expanded model for simulating urban growth. This model is the first to realize urban growth simulation based on CBR.

Point 2: Indexes listed in col.1 require more explanation because the description in col.2 "Process mode" is too general.

Response 2: We have modified Table 1 to more clearly describe the indexes in Col. 1, and the resulting table is shown below:

Table 1. Urban growth indicators and process mode.

Point 3: It is recommended to edit the text so that the references to Tables and Figures are closer to these objects.

Response 3: We edited the text so that Figures 2, 4, and 10 appear below the references paragraph.

We also tried to adjust Tables 6-7 and Figure 8, but since Table 6 takes up nearly a full page, we had to stick with the status quo.

Point 4: What is the advantage of the model presented in the paper over other models?

Response 4: CBR can even achieve quantitative analysis and prediction without needing to scrutinize the mechanism; this makes CBR particularly useful for geography processing, where the driving mechanism. Thus, CBR can effectively break through the bottleneck of knowledge acquisition present in traditional models, and therefore improve the reliability of geographic process simulations.

In this paper, we further compare the proposed CBR model with the current most commonly used CA model and find that the urban growth CBR model demonstrated higher accuracy, a simpler model construction, and a better ability to reflect trends in urban growth.

2004. Cheng; I. Masser, “Understanding Spatial and Temporal Processes of Urban Growth: Cellular Automata Modelling,” Environment and Planning B: Planning and Design, vol. 31, no. 2, pp. 167-194, 2004.
2005. K. Firozjaei; A. Sedighi; M. Argany; M. Jelokhani-Niaraki; J. J. Arsanjani, “A geographical direction-based approach for capturing the local variation of urban expansion in the application of CA-Markov model,” Cities, vol. 93, no. pp. 120-135, 2019.
2006. Liang; X. Liu; X. Li; Y. Chen; H. Tian; Y. Yao, “Delineating multi-scenario urban growth boundaries with a CA-based FLUS model and morphological method,” Landscape and Urban Planning, vol. 177, no. pp. 47-63, 2018.
2007. Tian; Y. Ouyang; Q. Quan; J. Wu, “Simulating spatiotemporal dynamics of urbanization with multi-agent systems—A case study of the Phoenix metropolitan region, USA,” Ecological Modelling, vol. 222, no. 5, pp. 1129-1138, 2011.
2008. Zhang; X. Jin; L. Wang; Y. Zhou; B. Shu, “Multi-agent based modeling of spatiotemporal dynamical urban growth in developing countries: simulating future scenarios of Lianyungang city, China,” Stochastic Environmental Research and Risk Assessment, vol. 29, no. 1, pp. 63-78, 2014.

He; Mai; Shen, “Delineation of Urban Growth Boundaries with SD and CLUE-s Models under Multi-Scenarios in Chengdu Metropolitan Area,” Sustainability, vol. 11, no. 21, pp. 5919, 2019.

2010. Luo; C. Yin; X. Chen; W. Xu; L. Lu, “Combining system dynamic model and CLUE-S model to improve land use scenario analyses at regional scale: A case study of Sangong watershed in Xinjiang, China,” Ecological Complexity, vol. 7, no. 2, pp. 198-207, 2010.
2011. Liu; Y. Yang; C. He; M. Tu, “Climate change will constrain the rapid urban expansion in drylands: A scenario analysis with the zoned Land Use Scenario Dynamics-urban model,” Sci Total Environ, vol. 651, no. Pt 2, pp. 2772-2786, 2019.
2012. Du; W. Wen; F. Cao; M. Ji, “A case-based reasoning approach for land use change prediction,” Expert Systems with Applications, vol. 37, no. 8, pp. 5745-5750, 2010.

Author Response File: Author Response.pdf

Reviewer 2 Report

I missed to understand the added value of this paper against previous (even, not cited) ones, such as http://dx.doi.org/10.1080/13658810600816870  I would enlarge significantly the cited literature and state much better the whole method (e.g. the CBR is practically a k-NN classification algorithm and that is not stated anywhere). I see that some work has been done for cross-validating against a 'classical' CA approach, but the whole experiment need more accuracy in description to be understandable and clearly posed. Value should be given to core originality aspects, that I honestly failed to catch.

I suspect that - as often - some off-the-shelf software modules have been used without a full comprehension of the algorithms behind, so I find the description of the whole CBR quite confused. I'm not an expert of urban spread analysis, but I found most of the article quite obscure and had to read other literature on CBR to fully understand the goal and contents. This is not a good sign, the paper could and should be written a lot more better and give evidence to originality aspects, that I truly struggle to understand.

Author Response

Response to Reviewer 2 Comments

Point 1: I missed to understand the added value of this paper against previous (even, not cited) ones, such as http://dx.doi.org/10.1080/13658810600816870  I would enlarge significantly the cited literature and state much better the whole method (e.g. the CBR is practically a k-NN classification algorithm and that is not stated anywhere). I see that some work has been done for cross-validating against a 'classical' CA approach, but the whole experiment need more accuracy in description to be understandable and clearly posed. Value should be given to core originality aspects, that I honestly failed to catch.

I suspect that - as often - some off-the-shelf software modules have been used without a full comprehension of the algorithms behind, so I find the description of the whole CBR quite confused. I'm not an expert of urban spread analysis, but I found most of the article quite obscure and had to read other literature on CBR to fully understand the goal and contents. This is not a good sign, the paper could and should be written a lot more better and give evidence to originality aspects, that I truly struggle to understand.

Response 1: We are very grateful for the reviewer's comments, and we are very sorry that the reviewer found this article difficult to understand. We want to respond to the reviewer's comments from three aspects.

Firstly, the reviewer thinks that “some off-the-shelf software modules have been used without a full comprehension of the algorithms behind.” In fact, we first constructed a CBR theory based urban growth simulation model from three aspects: case expression and collection, case retrieval and case constraint. Our using for reference the Du proposed CBR simulation for land use change ideas. We improved the case expression pattern and put forward the “initial state – geographical features – result” of case mode, proposing a strategy to introduce the "retrieval quantity" parameter and retrieve multiple most similar case, and gave a time factor control method based on demand constraints. The key ideas for these CBR simulations of urban growth were developed by us.

As for data processing by other software (eg. ArcGIS) in this paper, it is a common basic means to solve problems related to geographic information. These software only play the role of model realization, and do not affect the simulation process of CBR model of urban growth.

Secondly, in the process of constructing the model, this paper puts forward a large number of new terms, such as “initial state”, “geographical feature”, “result”, “geographical case”, “simulation case” and so on. These terms have been fully explained when they first appeared in the paper. For the common terms in the study of urban growth, this paper did not explain too much. The reason is that the model constructed in this paper is relatively complex, and we try our best to describe the model in more concise words, so as to make it easier for readers who do research related to urban growth to understand.

In fact, the CBR method adopted in this paper is not a common method to solve problems related to land simulation, and most scholars pay more attention to the CA model. In order to attract more relevant scholars to pay attention to CBR method, this paper deliberately compares CBR with CA, so as to attract scholars in urban growth simulation research to find the advantages of CBR method.

Finally, we did not quote the article[1] mentioned by the reviewer. In fact, we read this paper and thought that it was not related to our research. The CBR based CA model proposed by Li focuses on the mining of transformation rules by CBR, and the core method of its simulation is still CA, without in-depth discussion of the simulation method based on CBR reasoning mechanism. Therefore, the article[1] has little reference significance for the model construction in this paper. In addition, this paper has already been mentioned in the first chapter in paragraph 4: “Few studies have examined the spatial patterns or evolution processes of geographical phenomena using CBR. Among them, only one has proposed a CBR model for simulating land use changes.”

In addition, CBR understood as a K-NN classification algorithm is not accurate. The nearest neighbor method adopted in this paper is similar to K-NN classification algorithm. Although it is one of the commonly used methods in CBR retrieval, it is not the only retrieval method. Common retrieval methods also include K-D tree, footprint-based retrieval, validated retrieval and so on. In addition, case retrieval is only a link in CBR reasoning process, and it is also very important to express knowledge through case. Because CBR theory itself has a complex reasoning mechanism, this paper did not introduce CBR method in depth, but only introduced the basic ideas of CBR simulation in section 2.3: “Take the land patch as a case unit in order to solve the change result of land type in the new case; the old case can be retrieved from the known case base for the new case. When the old case that is most similar to the new case is retrieved, the results of the old case are applied to the new case. On this basis, this paper puts forward the following basic idea for applying CBR to urban growth simulation. Taking the land raster unit as the case, the case in the new period can retrieve the most similar cases in the old period, and then the land evolution type of the old case is taken as the urbanization result of the new case.”

2006. Li; X. Liu, “An extended cellular automaton using case‐based reasoning for simulating urban development in a large complex region,” International Journal of Geographical Information Science, vol. 20, no. 10, pp. 1109-1136, 2006.

Author Response File: Author Response.pdf

Reviewer 3 Report

The present article introduces CBR as a potential model for simulating urban growth using a study case from China.

The  results and the main conclusion(s) should be better emphasized in the abstract.

A very good aspect of this work is that the authors clearly explained the reasons for selecting the present case-study.

The article has a very good structure and coherence. However there are some aspects that should be corrected.

Please check the second equation because the symbols are not always  the same with the ones used in the following paragraphs.

Please improve the quality of  figure 6.

Please provide more information/ explanation on kappa coefficient and Figure of Merit (FoM). Which is their relevance? How they influence the results? etc.

Try to restructure the conclusions section in order to better emphasize the main conclusions of your work.

Overall, the article is very interesting and needs an extended minor revision.

Author Response

Response to Reviewer 3 Comments

Thanks for the reviewer's comments. We have responded to the comments in the following.

Point 1: The results and the main conclusion(s) should be better emphasized in the abstract.


Response 1: We have revised the abstract to further emphasize the results of the study. The detailed revised results are shown below:

Abstract: Developing urban growth models enables better understanding and planning of sustainable urban areas. Case-based reasoning (CBR), in which historical experience is used to solve problems, can be applied to the simulation of complex dynamic systems. However, when applying CBR to urban growth simulation, problems such as inaccurate case description, a single retrieval method, and the lack of a time control mechanism limit its application accuracy. In order to tackle these barriers, the study proposes a CBR model for simulating urban growth. This model includes three parts: (1) the case expression mode containing the “initial state-geographical feature-result” is proposed to adapt the case expression to the urban growth process; (2) in order to improve the reliability of results, we proposed a strategy to introduce the "retrieval quantity" parameter and retrieve multiple most similar cases; (3) a time factor control method based on demand constraints is given to improve the power of time control in the algorithm. Finally, Jixi City was taken as the study area for simulation, and when the “retrieval quantity” is 10, the simulation accuracy reaches 97.02%, Kappa is 85.51 and FoM is 0.1699. The results showed that the proposed method could accurately analyze urban growth.

Point 2: Please check the second equation because the symbols are not always the same with the ones used in the following paragraphs.

Response 2: We have modified the second equation, and the result is as follows:

                       (2)

Where l represents the initial land use type of the case, P is the simulated case, Q_i is the i^th geographical case, SIM_l (P, Q_i) is the similarity coefficient between the simulated and geographical cases under the land use type l, k is the driving index (geographic features) number, p_k is the k of the simulated case, q_k,i is the k of the geographical case, w_k is the weight assigned to k, and n is the index quantity.

Point 3: Please improve the quality of figure 6.

Response 3: We have improved the quality of Figure 6, and the new figure looks like this:

Point 4: Please provide more information/ explanation on kappa coefficient and Figure of Merit (FoM). Which is their relevance? How they influence the results? etc.

Response 4: Kappa coefficient and FOM index are commonly used indicators to evaluate the simulation of land use change/urban growth. There is not direct relevance between them, and each index can independently evaluate the simulation accuracy.

We have made a supplementary description in Section 3.1 of the original text, the details are shown below:

The simplest verification method for the model’s accuracy is to intuitively compare the simulation results with actual results. Through visual inspection, the simulated urban pattern of Jixi City in 2015 was compared with the actual urban pattern (Figure 7); the simulated pattern was basically similar to the real pattern. A confusion matrix of the concordance between the simulated and actual situations was then obtained to conduct quantitative analysis (Table 6), the results show that the simulation accuracy reaches more than 96%, and the kappa coefficient is also above 85% (Table 7). In addition, since Figure of Merit (FoM) is better than Kappa in the accuracy of evaluating the simulation changes[1], this paper adopts FoM to further evaluate the accuracy, the calculation formula is:

In this formula, A is the area of non-urban land that is transformed into urban land in the actual scenario, but not in the simulation scenario. B is the area of non-urban land transformed into urban land in both scenarios (the correctly transformed area). D is the area of non-urban land that is not transformed into urban land in the actual scenario but is in the simulation scenario.

Table 7 shows the calculated results of FoM, and it can be found that the FoM reaches more than 0.1600, reflecting the high accuracy of the simulation results. In order to further analyze the relationship between the simulation results and the retrieved quantity x, as can be seen from Figure 8, the simulation accuracy first increased and then decreased as x increased; it reached its highest value when x was ten.

Point 5: Try to restructure the conclusions section in order to better emphasize the main conclusions of your work.

Response 5: We restructured the conclusion section, and the revised results are as follows:

Simulating urban growth patterns is a necessary step for sustainable land-use management. The major contribution of this study is that it improves the CBR model traditionally used to simulate land use changes, thereby developing a CBR model for simulating urban growth. It redefined the case expression mode, developed the idea of case comprehensive retrieval, and proposed the introduction of a method to constrain the time factor.

This model was evaluated by taking Jixi City as the research area. It was found that when the parameter x=10, the simulation accuracy was 97.02%, kappa was 85.51, and FoM was 0.1699. The experimental results showed good simulation effects. In addition, the influences of different parameters and the validity of the model were discussed, and the effective categories for a comprehensive retrieval strategy were outlined. Compared with the CA model, the urban growth CBR model demonstrated higher accuracy, a simpler model construction, and a better ability to reflect trends in urban growth. However, the model requires a large amount of computation and has a slow running speed, we will mainly solve this problem in the future.

In summary, the proposed model provides a flexible, simple, and easy to understand method for studying the evolution mechanisms of urban patterns. The research results presented here can help to understand the evolution characteristics of urban spatial patterns, provide decision support for scientific urban regional planning, guide reasonable increases in construction land, and promote the sustainable and healthy development of cities.

2004. G. Pontius; D. Huffaker; K. Denman, “Useful techniques of validation for spatially explicit land-change models,” Ecological Modelling, vol. 179, no. 4, pp. 445-461, 2004.

Author Response File: Author Response.pdf