Scholarlink

This study investigates the use of genetic algorithm (GA) in optimizing neural network-based recommendation systems. In the initial part of the study, a comparative study on three variants of GA, viz. generational GA, steady-state GA, and Fuzzy GA, was carried out. Fuzzy GA was identified to be the most promising variant among its counterpart variants. For the implementation of Fuzzy GA, the Python Genetic Algorithm (PyGAD) and Distribute Evolutionary Algorithm in Python (DEAP) libraries were explored and the results are compared. The DEAP-based Fuzzy GA variant was selected for further experiments. In the later part of the study, the Fuzzy GA was integrated with a neural network-based recommendation system, which led to more advancements in the recommendations on a movie recommendation system. This integration simplified the process of optimizing the weights of the neural network as well as led to notable improvement in the recommendation system’s performance. This approach showed a significant increase in recall and F1-score which are considered consistent measures of performance. The effectiveness of the proposed recommendation system was confirmed by a comparison with benchmarking system in the literature.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Enhancing Movie Recommendation Systems Through Hybridization of Neural Networks and Genetic Algorithms

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Arvapalli Alekhya,
  • B. Priyadharshini,
  • N. Apoorva,
  • Gajula Sujan Sagar,
  • G. Jeyakumar

摘要

This study investigates the use of genetic algorithm (GA) in optimizing neural network-based recommendation systems. In the initial part of the study, a comparative study on three variants of GA, viz. generational GA, steady-state GA, and Fuzzy GA, was carried out. Fuzzy GA was identified to be the most promising variant among its counterpart variants. For the implementation of Fuzzy GA, the Python Genetic Algorithm (PyGAD) and Distribute Evolutionary Algorithm in Python (DEAP) libraries were explored and the results are compared. The DEAP-based Fuzzy GA variant was selected for further experiments. In the later part of the study, the Fuzzy GA was integrated with a neural network-based recommendation system, which led to more advancements in the recommendations on a movie recommendation system. This integration simplified the process of optimizing the weights of the neural network as well as led to notable improvement in the recommendation system’s performance. This approach showed a significant increase in recall and F1-score which are considered consistent measures of performance. The effectiveness of the proposed recommendation system was confirmed by a comparison with benchmarking system in the literature.