A Modified Incremental Learning Approach for Data Stream Classification

Data mining for data stream becomes important in academic areas. Due to large-scale data, people utilize incremental learning approach to handle the data. In this paper, a modified Support Vector Machine (SVM) incremental learning model is proposed. Through experiments of selecting kernel function for the SVM method, we optimize several parameters. Real network dataset is used in our experiments to verify the model´s feasibility and applicability. The experimental results show that the modified SVM incremental learning model can improve the accuracy of classification and increase performance.