Presenting a novel method based on collaborative filtering for nearest neighbor detection in recommender systems

Subject Areas : Intelligent Multimedia Processing and Communication Systems (IMPCS)

Mahdi  Bazargani ¹ (Assistant Professor)
Zeinab  Homayounpour ² (Faculty of Electrical and Computer Engineering, Islamic Azad University of Zanjan, Zanjan)

Received: 2021-07-04 Accepted : 2021-07-24 Published : 2020-09-22

Keywords: Nearest neighbor, Mean absolute error, collaborative filtering, Recommender Systems,

Abstract :

Recommendation systems propose specific items to users based on their interests by analysis the user data. The main goal of this analysis is extraction of each user pattern to predict the interested items. One of the main well-known methods in recommender systems is collaborative filtering in which similarity measures are utilized to detect similar users to a new user. The challenging issues related to collaborative filtering are similarity and neighborhood detection. In this paper, nearest neighbor (NN) algorithm is used to detect similar neighbors to a new user. The proposed model, which is inspired by user-item method, the score of items is calculated based on a distance metric and the nearest neighbor is selected. In the presented work, we detect similar users using user-item matrix and the Euclidean distance. The proposed method is evaluated on Movielens dataset which includes 1682 items and evaluation metrics such as Accuracy, Precision, Recall, F1-measure, Mean Absolute Error (MAE), and Root Mean Square Error (RMSE) are measured. MAE of the proposed method is 0.7351 which is less than Pearson and Cosine similarities, which demonstrates the superior performance of the proposed method in similarity detection and prediction.

References:

 

• [1] Duma, B. Twala, Optimising latent features using artificial immune system in collaborative filtering for recommender systems, Applied Soft Computing, Volume 71, pp. 183-198, 2018.


• [2] Jiang, S.C. Fang, Q. An, J.E. Lavery, A sub-one quasi-norm-based similarity measure for collaborative filtering in recommender systems, Information Sciences, Vol. 487, pp. 142-155, 2019.


• [3] Valcarce, A. Landin, J. Parapar, A. Barreiro, Collaborative filtering embeddings for memory-based recommender systems, Engineering Applications of Artificial Intelligence, Vol. 85, pp. 347-356, 2019.


• [4] Zhang, T. Gong, V.E. Lee, G. Zhao, Guangzhi Qu, Fast algorithms to evaluate collaborative filtering recommender systems, Knowledge-Based Systems, Vol. 96, pp. 96-103, 2016.


• [5] Mohammadpour, A.M. Bidgoli, R. Enayatifar, H.H.S. Javadi, Efficient clustering in collaborative filtering recommender system: Hybrid method based on genetic algorithm and gravitational emulation local search algorithm, Genomics, In press, corrected proof, Available online 3 January 2019


• [6] Parvin, P. Moradi, S. Esmaeili, TCFACO: Trust-aware collaborative filtering method based on ant colony optimization, Expert Systems with Applications, Vol. 118, pp. 152-168, 2019.


• [7] Kant, T. Mahara, V.K. Jain, D.K. Jain, A.K. Sangaiah, LeaderRank based k-means clustering initialization method for collaborative filtering, Computers & Electrical Engineering, Vol. 69, pp. 598-609, 2018.


• [8] Wasid, V. Kant, A Particle Swarm Approach to Collaborative Filtering based Recommender Systems through Fuzzy Features, Procedia Computer Science, Vol. 54, pp. 440-448, 2015.