基于遗传算法的灰自助法

发布时间：2018-05-16 01:10

本文选题：灰色预测模型 + 自助法　；参考：《重庆大学》2014年硕士论文

【摘要】：随着中国改革开放的发展，社会主义经济日趋完善，房地产业也呈现了一片欣欣向荣的景象。但是伴随而来的房地产投机现象也不断涌现，住宅房价增长迅猛，远远超出人民大众的购买力，楼市的泡沫经济已经严重影响国家经济的健康发展。国家和政府开始在2010年对住宅房价进行宏观调控，至此，中国房地产政策已由此前的支持转向抑制投机，为了遏制住宅房价过快上涨，国家还先后采取了土地、金融、税收等多种调控手段。至今，住宅房价迅猛增长的势头得到了有效的遏制。本文以调控开始有效的2011年至今的季度数据为依托去预测重庆以后季度的住宅房价格。然而如何建立一种可靠的预测模型，能够预测出参考价值高的住宅房价，仍是一个相对困难的问题。由于2011年至今的季度数据量小，所以它不能用通常经典的大样本预测方法去预测。本文首先就首先用一种可以模拟未知分布、再抽样统计的方法——自助法，对原始数据进行抽样，以达到增大样本容量的目的。然后，本文再结合灰色预测模型（灰色GM(1,1)模型），介绍了一种可以对小样本数据预测的灰自助模型。最后，阐述了一种带参数的灰自助模型，通过遗传算法（GA，GeneticAlgorithm）去修正参数得到一种能够使得预测值的全局平方误差最小的模型——基于遗传算法的灰自助模型（GA灰自助模型）。一般情况下，灰色预测模型和GA灰自助模型只能给出一个预测范围，然而这个预测范围往往太大，只能说明模型的正确性，不能给消费者一个具体的参考。但是消费者往往需要一个具体的参考，本文通过引入新的选择序列，，我们可以根据选择序列最后两个序列码来判定预测值的取值，进而得到预测值。
[Abstract]:With the development of China's reform and opening up, the socialist economy is becoming more and more perfect, and the real estate industry is also flourishing. But accompanied by the speculation in real estate is also emerging, housing prices are growing rapidly, far beyond the purchasing power of the public, the bubble economy of the property market has seriously affected the healthy development of the national economy. The state and the government began to macro-control housing prices in 2010. At this point, China's real estate policy has shifted from previous support to curbing speculation. In order to curb the excessive rise in housing prices, the state has also adopted land and finance successively. Taxation and other means of regulation and control. So far, the rapid growth of housing prices has been effectively contained. Based on the effective quarterly data from 2011 to present, this paper predicts the housing prices in Chongqing. However, how to establish a reliable prediction model to predict the housing prices with high reference value is still a relatively difficult problem. Because of the small amount of data in the quarter from 2011 to now, it can not be predicted by the classical large sample prediction method. In this paper, we first use a self-help method, which can simulate the unknown distribution and then sample statistics, to sample the raw data, so as to increase the sample size. Then, this paper introduces a grey self-help model which can be used to predict small sample data. Finally, a grey self-help model with parameters is presented. By modifying the parameters by genetic algorithm, a model with minimum global squared error of predicted value is obtained, which is a grey self-help model based on genetic algorithm (GA). In general, the grey prediction model and the GA grey self-help model can only give a prediction range, but this prediction range is often too large, which can only explain the correctness of the model and can not give consumers a specific reference. But consumers often need a specific reference. By introducing a new selection sequence, we can judge the predicted value according to the last two sequence codes of the selection sequence, and then get the predicted value.
【学位授予单位】：重庆大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F299.23;TP18

【参考文献】