基于支持向量回归的吉林省房地产价格预测模型研究

发布时间：2018-06-23 07:24

本文选题：聚类分析 + 房地产价格　；参考：《长春工业大学》2017年硕士论文

【摘要】：通过观察和研究房地产价格波动情况,人们能够对国家经济的发展情况和房地产市场的变化进行整体的把握,以便进行有针对性的投资或交易行为。近年来,随着我国房地产交易市场的活跃和快速升温,行业领域的潜在风险愈加的突出,如何使用合理有效的方法对房地产市场和房屋价格的波动情况进行反映,并及时提供正确的预警和提示变得非常重要。考虑到上述问题,本文从实际情况出发,对房地产市场价格的预测进行了研究。首先使用基于聚类分析的统计学方法对房地产价格产生影响的主要因素进行了分析和选择,这种指标选择方法有效解决了模型预测的代表性和全面性相互平衡的问题。在确定评价指标以后,我们基于国家统计年鉴和吉林省统计年鉴所公布的吉林省房地产数据,应用支持向量回归算法对吉林省房屋价格的变化规律进行了分析和预测。支持向量回归模型具有较好的统计理论基础和对少量数据情况下模型的较好拟合能力。然后,通过MATLAB软件编程语言实现了支持向量回归算法的构建、分析和预测。最后通过对比支持向量回归模型与RBF神经网络模型的预测误差,从结果上表明了支持向量回归模型的有效性和良好的预测精度。
[Abstract]:Through observing and studying the fluctuation of real estate price, people can grasp the development of national economy and the change of real estate market as a whole, in order to carry on the targeted investment or transaction behavior. In recent years, with the active and rapid warming of the real estate market in China, the potential risks in the field of the industry become increasingly prominent. How to use reasonable and effective methods to reflect the volatility of the real estate market and housing prices, It is important to provide the correct warning and warning in time. Considering the above problems, this paper studies the real estate market price forecasting from the actual situation. Firstly, the main factors influencing real estate prices are analyzed and selected by using the statistical method based on cluster analysis. This index selection method effectively solves the problems of representativeness and comprehensiveness of model prediction. Based on the real estate data of Jilin Province published in the National Statistical Yearbook and the Statistical Yearbook of Jilin Province, we apply the support vector regression algorithm to analyze and predict the changing law of housing prices in Jilin Province. The support vector regression model has a good statistical theoretical foundation and a good fit ability for the model with a small amount of data. Then, the support vector regression algorithm is constructed, analyzed and predicted by MATLAB software programming language. Finally, by comparing the prediction error between the support vector regression model and the RBF neural network model, the results show that the support vector regression model is effective and has good prediction accuracy.
【学位授予单位】：长春工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F299.23

【参考文献】