Title: Study on behaviour anomaly detection method of English online learning based on feature extraction

Authors: Feng Wei

Addresses: Yongcheng Vocational College, Yongcheng 476600, Henan Province, China

Abstract: There are many problems in abnormal detection of online English learning behaviour, such as large error and high detection time. Therefore, a detection method based on feature extraction is proposed. Firstly, frequent pattern mining method is used to collect learners' behaviour data, and the data is collected and preprocessed. Then, the classification constraints are set by support vector machine to complete the data classification. Finally, the sequence minimum eigenvalue method is used to train the abnormal data, extract the high frequency features of the abnormal data, establish the anomaly detection model, and realise the anomaly detection. Experimental results show that the highest detection error of this method is 1.2%, and the highest time cost is 1.8 s. Therefore, this method can effectively reduce the detection error and time cost, and is feasible.

Keywords: feature extraction; English online learning behaviour; anomaly detection; threshold; K-means clustering; Lagrange function.

DOI: 10.1504/IJRIS.2023.128372

International Journal of Reasoning-based Intelligent Systems, 2023 Vol.15 No.1, pp.41 - 47

Received: 25 Feb 2022
Accepted: 28 Apr 2022
Published online: 18 Jan 2023 *