Title: Abnormal data classification of social media based on support vector machine

Authors: Kangyi Wang

Addresses: Department of Computer Science, Changzhi University, Changzhi 046011, China

Abstract: Dataset training is ignored in abnormal data classification of network social media, which leads to large errors in data classification results. This paper proposes a classification method for abnormal data of social media based on support vector machine (SVM). Online social media data through the curvature features are divided into two different regions, and with the aid of different data denoising filtering methods, completes the pre-processing of the network social media data. By using genetic algorithm to improve K means clustering algorithm, then by completing the training set of abnormal data through the SVM classifier, selecting samples closest to the hyperplane as negative samples, and training SVM at the same time, completes the abnormal data classification of network social media. The results show that the proposed method has the highest error rate of about 1.25% in the classification of abnormal data in network social media, and the data classification can be completed in a relatively short time.

Keywords: support vector machine; SVM; network social media; classification of abnormal data; noise removal and filtering; negative sample.

DOI: 10.1504/IJWBC.2022.125504

International Journal of Web Based Communities, 2022 Vol.18 No.3/4, pp.249 - 261

Received: 08 Jun 2021
Accepted: 05 Nov 2021
Published online: 12 Sep 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article