[data analysis] realize SVDD decision boundary visualization based on MATLAB

1 Content introduction

In the real world , Everything has its characteristics , Such characteristics are more or less 、 Or important or unimportant . People can determine the classification of things by their characteristics , But when things have many characteristics , If people rely on traditional methods to classify things, it will be time-consuming and labor-consuming , And the accuracy of classification is not high . And classification as a prediction model , If the accuracy of classification is low or the time is long , Then this prediction will become worthless . Therefore, people have proposed various classification models to predict things , Among them, support vector machine and support vector description data have certain advantages in predicting high-dimensional data , And according to different requirements , The improvement of these two algorithms has been applied to many fields in real life . First , This paper studies the background and theory of support vector machine in Data Mining Classification Algorithm , Analyzed and summarized SVM Research status of various improvement methods .​

2 Simulation code

% DESCRIPTION% Visualize decision boundaries of SVDD (libsvm-3.23)​clcclose alladdpath(genpath(pwd))​[traindata, testdata, trainlabel, testlabel] = prepareData;​c = 0.8;  % trade-off parameterg = 1/16; % kernel width​% Train SVDD hyperspherecmd = ['-s 5 -t 2 ', '-c ', num2str(c), ' -g ', num2str(g), ' -q'];model = libsvmtrain(trainlabel, traindata, cmd);​% Predict testing data[predictedlabel_test, acc, ~] = libsvmpredict(testlabel,testdata, model); ​% Decision Boundary[rho, X1, X2]  = decision_boundary(model, traindata);​% Visualizationpara = [c, g, acc(1)];visualization(model, traindata, testdata, rho, X1, X2, para)​​​​​​

3 Running results

4 reference

[1] Lichuanliang . be based on SVDD And parameter identification of analog circuit fault diagnosis method [D]. Nanjing University of Aeronautics and Astronautics .

[2] Shotto . An improved support vector data description algorithm [D]. Harbin Engineering University , 2013.

About bloggers ： Good at intelligent optimization algorithms 、 Neural networks predict 、 signal processing 、 Cellular automata 、 The image processing 、 Path planning 、 UAV and other fields Matlab Simulation , relevant matlab Code problems can be exchanged by private letter .

Some theories cite network literature , If there is infringement, contact the blogger to delete .