河北省科学院学报

本研究针对互联网上影视作品数量的指数级增长，提出一种基于Spark分布式框架的电影推荐系统，以解决日益严重的电影信息过载问题。系统利用Spark平台作为计算框架，通过ALS协同过滤算法构建一个电影推荐模型，并使用MovieLens数据集进行了训练。实验结果表明，该系统可根据用户历史数据和实时动作快速准确预测用户个人喜好，同时也解决了传统推荐系统的数据稀疏性和冷启动问题。

To address the exponential growth of online film and television content, this study proposes a Spark-based distributed movie recommendation system to solve the increasingly severe problem of movie information overload. The system adopts Spark as the computing framework, constructing a movie recommendation model through the Alternating Least Squares(ALS) collaborative filtering algorithm, and is trained on the MovieLens dataset. Experimental results demonstrate that the system can accurately and rapidly predict individual users′ preferences based on their historical data and real-time interactions. Furthermore, it effectively resolves data sparsity and cold-start issues in traditional recommendation systems.