Hybrid Software Obsolescence Evaluation Model Based on PCA-SVM-GridSearchCV

Abstract

For the sake of establishing an accurate software obsolescence evaluation model scientifically and systematically, this paper points out the mechanism and definition of software obsolescence. By the way of analyzing software obsolescence mechanism, 20 software obsolescence correlative characteristic parameters are collected, preprocessed and scaled by RobustScaler, and then PCA is used to extract feature and reduce dimension, eliminate the noise value in the feature data and select important software obsolete feature data. Use GridSearchCV class to optimize support vector machine parameters and build SVM classification model, use confusion matrix accuracy to evaluate the machine model, finally the example verify the evaluation model to be believable and effective.