Simulating and predicting the urban land use change can provide deeper spatial insights into dynamics and sustainable developments of urban planning. This research takes the Shanghai-Hangzhou Bay (SHB) agglomeration as a study area and selects natural, economic, social, and policy variables as restraint conditions. A cellular automaton (CA) model and the naive Bayes-cellular automaton (NB-CA) model are employed and compared to simulate the built-up land in SHB. Results show that the NB-CA model greatly improves the simulation accuracy of built-up land compared to CA model. Specifically, the simulation accuracy of the NB-CA model is 14.68%, 14.03%, 7.43%, 6.00%, 5.32%, and 2.65% higher than that of the traditional CA model in Shanghai, Hangzhou, Huzhou, Jiaxing, Ningbo, and Shaoxing, respectively. Among the four restraint conditions, the least influential variable is the natural variable and the most influential is the policy variable in Shanghai, Ningbo, and Shaoxing and the social variable in Hangzhou, Huzhou, and Jiaxing. It is the first usage of naive Bayes and CA to simulate built-up expansion and this new combination method highlights the improvement of simulation accuracy. The naive Bayes technology implies that government policy is an unstable factor that can influence the simulation of built-up land change. The methodology will be applicable to other regions experiencing rapid built-up land expansion under government policy.

中文翻译：

---

结合朴素贝叶斯和元胞自动机模型的建成区扩张模拟——以沪杭湾群落为例

模拟和预测城市土地利用变化可以为城市规划的动态和可持续发展提供更深入的空间洞察。本研究以沪杭湾（SHB）群落为研究区域，选取自然、经济、社会和政策变量作为约束条件。采用元胞自动机 (CA) 模型和朴素贝叶斯元胞自动机 (NB-CA) 模型进行比较，以模拟 SHB 的建成区。结果表明，与CA模型相比，NB-CA模型大大提高了建设用地的模拟精度。具体而言，NB-CA模型的仿真精度比上海、杭州、湖州、嘉兴、宁波等地的传统CA模型分别提高了14.68%、14.03%、7.43%、6.00%、5.32%和2.65%。绍兴分别。在四种约束条件中，影响最小的变量是自然变量，影响最大的是上海、宁波和绍兴的政策变量和杭州、湖州和嘉兴的社会变量。这是首次使用朴素贝叶斯和 CA 来模拟组合扩展，这种新的组合方法突出了模拟精度的提高。朴素贝叶斯技术意味着政府政策是一个不稳定的因素，可以影响建成用地变化的模拟。该方法将适用于在政府政策下经历快速建设用地扩张的其他地区。这是首次使用朴素贝叶斯和 CA 来模拟组合扩展，这种新的组合方法突出了模拟精度的提高。朴素贝叶斯技术意味着政府政策是一个不稳定的因素，可以影响建成用地变化的模拟。该方法将适用于在政府政策下经历快速建设用地扩张的其他地区。这是首次使用朴素贝叶斯和 CA 来模拟组合扩展，这种新的组合方法突出了模拟精度的提高。朴素贝叶斯技术意味着政府政策是一个不稳定的因素，可以影响建成用地变化的模拟。该方法将适用于在政府政策下经历快速建设用地扩张的其他地区。