Article: Cloud-based user behaviour analysis and personalised recommendation of sports teaching system based on big data analysis Journal: International Journal of Continuing Engineering Education and Life-Long Learning (IJCEELL) 2026 Vol.36 No.7 pp.94 - 116 Abstract: The fragmentation and dynamic changes of user behaviour data in the current cloud basketball teaching system lead to delayed recommendation content and low matching degree. This paper proposes a personalised recommendation method based on big data analysis and deep learning. It combines the long short-term memory (LSTM) network to extract training time series features, and integrates the collaborative filtering algorithm to optimise personalised recommendations to accurately identify users' training rhythm and technical shortcomings. Experiments show that after 50 training cycles, the model's root mean square error (RMSE) reached 0.36 and the mean absolute error (MAE) was 0.33. After the personalised recommendation system, the user training completion rate was improved, ranging from 20% to 30%. This method can effectively improve the intelligence level of basketball teaching system and promote the realisation of personalised learning. Inderscience Publishers - linking academia, business and industry through research

Title: Cloud-based user behaviour analysis and personalised recommendation of sports teaching system based on big data analysis

Authors: Guohua Li

Addresses: Public Course Teaching Department, Linfen Vocational and Technical College, Linfen 041000, Shanxi, China

Abstract: The fragmentation and dynamic changes of user behaviour data in the current cloud basketball teaching system lead to delayed recommendation content and low matching degree. This paper proposes a personalised recommendation method based on big data analysis and deep learning. It combines the long short-term memory (LSTM) network to extract training time series features, and integrates the collaborative filtering algorithm to optimise personalised recommendations to accurately identify users' training rhythm and technical shortcomings. Experiments show that after 50 training cycles, the model's root mean square error (RMSE) reached 0.36 and the mean absolute error (MAE) was 0.33. After the personalised recommendation system, the user training completion rate was improved, ranging from 20% to 30%. This method can effectively improve the intelligence level of basketball teaching system and promote the realisation of personalised learning.

Keywords: physical education system; personalised recommendations; big data analysis; long short-term memory network; LSTM; collaborative filtering; mean absolute error; MAE.

DOI: 10.1504/IJCEELL.2026.151825

International Journal of Continuing Engineering Education and Life-Long Learning, 2026 Vol.36 No.7, pp.94 - 116

Received: 13 Mar 2025
Accepted: 11 Sep 2025
Published online: 20 Feb 2026 *

Title: Cloud-based user behaviour analysis and personalised recommendation of sports teaching system based on big data analysis

Keep up-to-date