Prediction Of Landslide Displacement Series By The Support Vector Machine

With the country economy developing and enlargement of rock and soil engineering, the langslide problems lead to more and more recognition by people. Landslide has become serious geological hazardousness in our living and production after earthquake volcano that complex mechanism, influence many factors: on the base of landslide destruction mode, transmutation evolution process of complex levity. It is very difficult to establish precise mathematic model to simulate landslide stability and to establish prediction simulation to predict time and dimension evolution. The suggestion and advance of intelligent rock mechanics provide the landslide prediction with a new track. With the support vector machine (SVM) and the genetic algorithm (GA) being combined, it has better adaptability for the question that is difficult to be described by mathematic model and has also widely recent advances. The evolution SVM consists and the landslide displacement is predict with the the genetic algorithm and the support vector machine combined, in the basis of people beor. The primary tasks are stated as followed:(1) The parameters of the support vector machine are searched speedily and the representational samples are speedily constituted, such as the core function and punishing parameters, which makes use of the best characteristic, for example, little samples, high dimensions, nonlinearity, and make use of the improved genetic algorithm.(2) The predicting module of the visiual and predicting landslide system is exploit vc++ developed tool with the theory of the support vector machines and the genetic algorithm.(3) With the monitoring displacement data of Bachimen of Fujian and Longtan of Guangxi analysed thoroughly, the studying samples are consited according the corresponding displacement data.It is that the landslide displacement model of the one factor time series are established, and the displacement of the Bachimen landslid are predicted. The studying samples are consisted with NN for the same data, the prediction...