Movie recommendation method using improved multi-objective genetic algorithm based on grid and difference substitution

Abstract

The invention relates to a movie recommendation method using an improved multi-objective genetic algorithm based on grid and difference substitution, provides an improved algorithm GDNSGA (Gradient Descent Non-dominated Sorting Genetic Algorithm)-II for defects of distribution and convergence in an NSGA (Non-dominated Sorting Genetic Algorithm)-II, and can be used for solving the problem of multi-objective combinational optimization. In the algorithm, an initialization population is designed by use of a multi-objective grid partition mode, thus the loss of distribution caused by non-uniform individuals is avoided; a population evolutionary process is maintained by use of clustering selection and difference substitution operators; and the appropriate number of inferior individuals is selected for local search, and the convergence and distribution of the population are kept. In combination with exploring of user behaviors and movie properties, the algorithm is used for the practical problem of personalized recommendation of a movie; through test comparison between the algorithm and an existing algorithm, the universality and effectiveness of the algorithm are explained, an excellent recommendation result is obtained; and the recommended F harmonic rate, diversity and novelty are improved, and an abundant recommended scheme combination is provided, thereby being beneficial to fully exploring the interest of a user, and providing a reliable recommendation service.

Description

technical field [0001] The invention belongs to the technical field of personalized recommendation. Using the improved multi-objective genetic algorithm GDNSGA-II algorithm for the shortcomings of the NSGA-II algorithm (specifically involving the NSGA-II algorithm, multi-objective grid division method, cluster pruning and difference replacement operator) to realize the personalization of movies recommend. Background technique [0002] With the popularization of Internet technology and the rapid development of modern e-commerce, the amount of resources flooding the Internet is increasing exponentially. The simultaneous presentation of a large amount of information often makes users feel at a loss, and it is difficult to find the resources they are really interested in, resulting in the so-called "information explosion" and "information overload" phenomena. Search engines and information retrieval technologies emerged to alleviate the problem of information overload. In tod...

AI Technical Summary

Problems solved by technology

NSGA-II is the most widely used multi-objective evolutionary algorithm. The characteristic of this algorithm is to determine the individual fitness value according to the Pareto dominance relationship and density information among individuals. However, there are distribution and convergence maintenance in such a fitness calculation method. improper defect

Wen Shihua et al. improved NSGA-II by reta

Images