The Forecast Of Bearing Value Of Foundation Soil Based On SVM

Support Vector Machine is a kind of machine learning arithmetic based on Statistical Learning Theory. It is a learning machine aiming at a limited sample and can reach a very good classified effect under the condition of less training samples. Its foundational theory is the theory of VC Dimension and Structural Risk Minimization on the statistics theory; It translates the classified problems into solving a convex second programming problems in the end; It converts the real matter into character space of high dimension through nonlinear mapping, in which linearity decision-making function can come into none-linearity decision-making function in original space, The problem"The Dimension Disaster"has been skillfully solved, and promised that there is better promoting ability. Moreover, the complexity of this method is not related to the dimension of the sample.The text is based upon the theory of SVM and SV regressive Machine. And it is also a Return Arithmetic Method--Data Preprocessing --SV regressive Machine which is put forward aiming at solving the real problems, that is, a major data --the forecast of bearing value which is used for designing a building groundwork.When gain sampled data (i.e. foundation soil physical data), because gathered is not very widespread, reasons and so on physical test to create certain error, caused the data not to be unstable. Then in classification time, the classified effect is not good. In order to achieve the good classified effect, therefore we carry on processing to the data are the data preprocessing first, uses the three dimensional filter to be possible to produce certain new samples. Suppose altogether may obtain l new samples. Then carries on the return with these new samples, available following return algorithm: SV regressive Machine Arithmetic Methods with the rigidityεor single parameter restraint SV regressive Machine Arithmetic Methods and so on, thus obtains Machine of classify.(Ⅰ)SV Return Machine Arithmetic Methods with the rigidityε:(1)Given Training Set:(2)Select proper precision parameterεand the core K ( x , x′) ;(3)Construct and try to uncoil the optimization problems: Get the best solution:(4)Or select the positive component ofα:αj>0 ,calculate according to Or select the positive component ofα? :α?j > 0, calculate according to(5)Construct the decision-making function( II ) SV Return Machine Arithmetic Methods under the restriction of single parameter(1) Given Training Set: T= thereinto,(2)Select proper positive parameterεand C; Select proper core: K ( x , x′) ;(3)Construct and try to uncoil the optimization problems: mαGet the best solution:(4)Or the choice satisfiesαi +αi? < C andαj>0 calculate according to this:Or the choice satisfiesαi +αi? < C andα?j>0 ,calculate according to this(5)Construct the decision-making functionIn the experiment sect foundation soil is used as the physics data, and the bearing value of foundation can be obtained by the SV regressive Machine Arithmetic Methods.The physical dada 52 entries of the foundation soil are obtained through drilling, collecting and disposing on the spot. The SV regressive Machine Arithmetic Methods with the rigidityεor SV regressive Machine Arithmetic Methods under the restriction of single parameter and the Arithmetic Method of Data Preprocessing --- SV regressive Machine are early or later utilized and come true through the Matlab 6.5 programming. The disposal result in the real data shows:(1) The Arithmetic Method of Data Preprocessing --- SV regressive Machine is effective and feasible. The arithmetic method can increase the stability of the sample, play down the individual number of the sample. At a certain extent, it precedes The SV regressive Machine Arithmetic Methods with the rigidityεor SV regressive Machine Arithmetic Methods under the restriction of single parameter.(2)Use the Method of Data Preprocessing --- SV regressive Machine, Thus through to foundation soil physics data processing, it can accurate separate the soil layer obtains the earth bearing value.It has the most important significance to measure the bearing capacity accurately. Its significance is if the bearing capacity is measured accurately, it can make the foundation of the building both firm and low in the cost while designing it; Conversely, if the measurement of the bearing capacity is not accurate, it will not noly probably make the foundation of the building waste lots of building materials due to the unnecessary hugeness, but also mayby make the building collapse terriblely because it is not firm.