紀要論文

鑑賞履歴に基づいた映画推薦アルゴリズム

Recommendation Algorithms Based on Behavioral History.

jpn

24368083

https://doi.org/10.15002/00026387

法政大学大学院理工学研究科

法政大学大学院紀要. 理工学研究科編

64

1

3

2023-03-24

Version of Record

With the rapid popularity of Internet and mobile Internet, the number of movie entertainment information on the Web is quite huge, and it is increasingly difficult for people to obtain information about movies of interest. In this study, we propose a personalized recommendation strategy based on the sequence of user playing behavior for the personalized recommendation problem of movie websites. The strategy analyzes the user playing video behavior data by Word2vec, a deep neural network word vector model, maps the videos into equal-dimensional feature vectors, calculates the similarity of movies, and uses it as the basis of recommendation to generate a recommendation list to users. In addition, to improve recommendation accuracy, Word2vec parameter settings were considered, and a pre-processing method for history sequences was proposed.

Article