Research On Incremental Classification Method Of Least Squares Support Vector Machine

Support vector machine(SVM)is a very powerful machine learning model that can perform linear or non-linear classification and has been widely used in various fields.However,in practical problems,various uncertain factors in the external environment will impact on the successful achievement of objectives.A large number of scholars have tried to improve the traditional support vector machine.In recent years,least squares support vector machine and twin support vector machine are researched hotspots,but their classification on high-dimensional problems and outlier problems is less than satisfactory.This paper introduces a progressive idea to improve the support vector machine,the specific research content is as follows:(1)In order to reduce the effect of the external environment,combining human cognitive rules and the classification principle of support vector machine,a three-way decisions based incremental support vector machine classification algorithm was proposed.Firstly,the positive region,negative region and boundary region of the sample are defined in the SVM classification.Secondly,compute the corresponding threshold set in the boundary domain according to the threshold.Then make the further decision in the boundary domain.Finally,the validity and feasibility of the algorithm are verified by UCI the data.(2)The least squares support vector machine(LSSVM)is difficult to deal with the high dimensional data problems.This paper proposes a feature selection algorithm based on L1 norm least square support vector machine(LSSVM),and gives its proof.The algorithm uses L1 norm to control the components of LSSVM to realize iterative feature selection,and the effectiveness of the proposed method is proved by UCI data experiments.(3)Compared with the traditional support vector machines,fuzzy least squares twin support vector machine(FLSTSVM)has the advantages of fast training speed and strong classification ability.However,this model ignores the position characteristics of sample points in the feature space,which may lead to wrong calculation when outliers are treated as support vectors.So,we propose an intuitionistic fuzzy least square twin support vector machine,which introduces a scoring function to effectively reduce the influence of noise and outliers.