Application Of Soil Slope Stability Evaluation Based On Decision Tree Algorithm

Slope stability’s analysis and evaluation is a important research area in geotechnical engineering for a long time. From limit equilibrium method to uncertainty analysis method that widely used today, a large number of studies have been conducted by researchers. Slope stability and related problems researched by different methods in different angels. For example, the reliability analysis method analyze the slope’s stability and failure probability in the angle of Statistics; the limit equilibrium method build equation of static equilibrium to solve slope stability statuses in the angle of slope’s limit states; the fuzzy comprehensive evaluation apply fuzzy mathematics to analyze the slope stability in the angle of information’s randomness and fuzziness. But different method has its own limitations. In fact, slope engineering is a complex huge system which is constantly changing and open. Deterministic calculation can’t summarize its complexity very well, moreover, most of undeterministic calculation methods, such as LVQ neural network, the accuracy of prediction of network always related to the differences of the data. Therefore, build a slope stability evaluation model which is low requires to data and good general performance will be significant.According to the feature of soil slope stability which is easy measurement and highly non-linearity,this article introduce the decision tree algorithm to soil slope stability evaluation in order to form a comprehensive evaluation method which is based on decision tree. It includes:The meaning of this article are illustrated briefly, meanwhile, the research of soil slope stability at home and abroad are summarized. The common soil slope stability’ evaluation methodology has been described systematically. At the same time, these methods’ advantages and disadvantages has been summarized and analyzed.Decision tree’s wide application is introduced briefly. The common decision tree classification algorithm has been described systematically. At the same time, these algorithms’ scope of application has been pointed out by comparative analysis. The feasibility of decision tree classifier apply in soil slope stability evaluation has been demonstrated.Decision tree algorithm has been applied in soil slope stability evaluation for the first time. This article builds the model of soil slope stability evaluation based on C4.5 decision tree algorithm. In order to compared and analyzed by several models, this article build the model of soil slope stability evaluation based on BP neural network and the model of soil slope stability evaluation based on LVQ neural network. The advantage of applying decision tree algorithm in soil slope stability evaluation has been demonstrated.