| Yunnan province is located in the southwest of Chinese, the province’s total population of more than 47 million, the total area of about 39 thousand square kilometers, the province has 8city, 8 National Minority Autonomous Prefecture, China connecting Southeast Asian countries bridgehead. Use of province of unique tourism resources, agricultural resources and mineral resources and Southeast Asian countries trade, Yunnan Province, the speed of economic development in recent years has come out in front, but at the same time, Yunnan Province geological disasters of debris flow, landslides and other times and scale are growing continuously, at the same time the economic development of water by the debris flow, landslides are more and more big. Yunnan province suffered from frequent geological disasters because of its mountainous landform in the Yunnan- Guizhou Plateau, and the province from the north to the South as the ladder fell, the highest and lowest height difference of up to more than 6000 m. In the rain in flood season, Yunnan province are often affected by rainfall impact large areas, and in recent years because of excessive exploitation of underground resource over exploitation and forest resources, natural factors, coupled with the influence of human factors causing landslides, debris flow and other geological disasters frequent. From 1953 to 2014 a total of because of debris flow, landslide caused the death, the number of missing up to more than 1 people, the number of injuries due to disasters is also up to 22 return, the direct economic loss of eight billion RMB. Such as the calculation of the indirect economic loss caused by the amount, will be amazing tens of billions of RMB. In this paper, through the collection of the landslides in Yunnan Province in 2011, 1998 to the various counties of the relevant data of landslide disaster, the coefficients to describe the relevant economic loss as the explained variable, disaster grade, affected county when years of average annual precipitation, the affected district population density by such disasters occurred in the disaster area, a total of three years of four times as the explanatory variables of debris flow, land slide,do an empirical analysis on the economic effect of Yunnan province.First through the analysis of the explanatory variable and the explanatory variable coefficient of disaster affected areas county disaster grade, year average annual precipitation, the affected counties by population density, the relationship between the number of disaster area county before three years total such disasters by using conventional statistical model, found the disaster affected districts of the coefficient and a certain population density the negative correlation relationship with the outside, no other explanatory variables, and from the R-squared and establishment of Q-Q map analysis of the data is difficult to use the traditional method of statistical modeling, so the use of machine learning.Secondly, the traditional statistical models are often in the face of complex data be at a loss what to do, it is difficult to use the distribution function of a certain to make assumptions. Then without any assumptions and algorithm based machine learning method can reflect the unique advantage. The decision tree using machine learning methods in regression, artificial neural network regression, support vector regression, Adaboost regression, random forest return these common machine learning method to analyze the data, and accurately by using the letter of 50 percent off cross validation proved that machine learning methods.Finally, the analysis results show that random forest is the optimal use of the model, carries on the reasonable prediction based on the disaster by the coefficient of model, to the original disaster prevention, disaster mitigation system in Yunnan Province, puts forward some scientific suggestions. |
PDF Full Text Request