Research on Vanadium and Titanium Culture Knowledge Recommendation Model Based on Collaborative Filtering Algorithm on the Spark Platform

With the deepening application of vanadium and titanium culture knowledge base, the data scale has expanded dramatically, making it increasingly difficult for users to find the required information in the huge knowledge base. However, the current knowledge base cannot recommend the knowledge that users are interested in. This paper is based on big data processing technologies such as Spark, Kafka, and Hive to collect log data for cleaning knowledge bases, and adopts the improved cosine similarity and collaborative filtering algorithms based on user preferences to recommend personalized knowledge for users. This system can efficiently process offline and real-time data, and generate TOP-N recommendation results. The experimental results show that the system has significant improvement in data processing rate, recommendation accuracy, and system stability compared with traditional methods.