Title: Personalised resource recommendation method for collaborative tagging system based on machine learning

Authors: Xiaofei Liu

Addresses: School of Artificial Intelligence, The Open University of Guangdong, GuangZhou 510091, China

Abstract: In order to overcome the low feasibility of traditional resource recommendation methods, this paper proposes a personalised resource recommendation method based on machine learning. Firstly, the user-based collaborative filtering algorithm is used to calculate user personalised similarity, and then content-based collaborative filtering algorithm is used to calculate resource content similarity through cosine similarity. Combined with user similarity and resource content similarity, a hybrid computing model of resource similarity is established, and personalised recommendation is realised through statistical machine learning. The experimental results show that: the F-measure value of the method can reach 0.97, the coverage rate is maintained above 50%, the popularity is above 0.8, and the MAE value is always the minimum, and its precision is always higher than that of the contrast method. It shows that the proposed method can effectively improve the precision and feasibility of personalised recommendation results.

Keywords: machine learning; collaborative tagging system; personalised resource recommendation; similarity.

DOI: 10.1504/IJISE.2022.126023

International Journal of Industrial and Systems Engineering, 2022 Vol.42 No.1, pp.1 - 19

Received: 14 Sep 2020
Accepted: 23 Nov 2020
Published online: 07 Oct 2022 *

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