Enhancing Cross-Domain Recommendation System using a Novel hybrid NLP Based Text Vectorization and Unsupervised Machine Learning Model

Recommendation systems play a pivotal role in enhancing user experiences in various domains. Cross-domain recommendations, where preferences and behaviours from one domain are leveraged to make recommendations in another, are of particular interest. This research paper presents approach to enhance cross-domain recommendations through a hybrid system that combines natural language processing text vectorization and k-means based modelling. The methodology in this paper involves converting movie titles into high-dimensional vectors, which capture semantic information related to the films. These vectors are then integrated with a k-means-based recommendation model, which allows to recommend movies as well as books based on movie titles. To evaluate the effectiveness of the mentioned approach, few comprehensive experiments on real-world data was conducted. The results demonstrate a significant improvement in recommendation accuracy and the ability to bridge the gap between seemingly disparate domains. There is a discussion on the implications of the findings and the potential for broader applications. This research paper contributes to the ongoing efforts to improve cross-domain recommendation systems and showcases the power of hybrid techniques.