Design and Implementation of a Movie Recommendation System Based on Hybrid Recommendation Algorithms

Xiaowei Liu, Zhiwen Yu

Keywords: movie recommendation;collaborative filtering;hybrid recommendation

Abstract:

With the rapid expansion of online multimedia platforms, the ability to deliver personalized content becomes increasingly critical. Movie recommendation systems play a central role in helping users navigate large-scale content repositories and identify items aligned with their preferences. However, traditional single-method approaches such as collaborative filtering often struggle with data sparsity, cold-start issues, and limited robustness. To address these challenges, this paper presents a hybrid recommendation framework that integrates User-based Collaborative Filtering (UserCF) with Singular Value Decomposition (SVD). By combining neighborhood-level similarity signals with latent feature representations, the proposed model aims to improve the overall stability and applicability of recommendation outcomes. The hybrid model is trained on the MovieLens dataset and employs an adaptive weighting strategy to dynamically fuse the outputs of the two components.
To demonstrate the framework’s practical viability, we develop a fully functional movie recommendation system using Django, supported by MySQL for data management and Bootstrap for a responsive user interface. In addition, detailed movie metadata corresponding to MovieLens entries is scraped from the Douban platform to enrich the system’s information content. Experimental observations and system deployment results indicate that the hybrid method operates reliably in real-world settings and provides a smooth user experience with consistent recommendation quality. Overall, the proposed framework bridges algorithmic design and practical deployment, offering a feasible solution for building personalized recommendation services in modern online media environments.

Submission Number: 59