Turkish Journal of Computer and Mathematics Education
Journal license

Journal

Turkish Journal of Computer and Mathematics Education


Volume
& Issue

Volume 12, Issue 2


Published
on

April 5, 2021


Pages

3313-3319


DOI

Article

A Matrix factorization technique using parameter tuning of singular value decomposition for Recommender Systems


Authors

Geluvaraj B Affiliation:
Research Scholar, Bangalore
and Meenatchi Sundaram Affiliation:
Associate Professor, Garden City University


Abstract

In this article we explained the concepts of SVD and algorithm evolution. MF technique and the working of it with computational formulas. PCA with step-by-step approach with example and A novel approach of Hyper SVD and How to fine tune it and pseudocode of the Hyper SVD with the Experimental setup using SurpriseLib and computing RMSE and MSE for the accuracy purpose and solving with the real time example which solves the cold start hassle also together and it can be seen that comparison of SVD and Hyper SVD and Random algorithm is done and types of Movies they recommended . There is far more difference between the results of the both algorithms and movie recommendations as per the results Hyper SVD is flexible and efficient and superior compared to other algorithms.


Keywords

Matrix Factorization (MF), Singular Value Decomposition (SVD), Principal Component Analysis (PCA), Machine Learning (ML)


Citation

B, G. & Sundaram, M. (2021). A matrix factorization technique using parameter tuning of singular value decomposition for recommender systems. Turkish Journal of Computer and Mathematics Education, 12(2), 3313–3319.

Published by: Engineering Journals

Engineering Journals Logo