Movie Recommender System with Visualized Embeddings
Sreenivas Jayanth Yadhati, S. Vikyath, Vishnu Priya Maddela, Vishvesh Pathak, C. P. Prathibhamol · 2021
Recommender Systems are a fundamental building block of E-commerce and Online Streaming services. With the advent and subsequent globalization of the internet and data, integration of Machine Learning and Deep Learning Algorithms into these online services has happened. In this approach combination, There are be methods implemented, like providing input-based embeddings for users and items (which are generated via models obtained from a Standard Deep Neural Network), to get results. By utilizing the t-Distributed stochastic Neighbor Embedding Tool, visualizing this complex data in lower dimensions has been done. KNN has been implemented to search for similar embeddings. Thus a combination of methods has been utilized to both recommend movies and visualize the generated embeddings, along with being applied on the movielens 100K dataset. Approach mayor may not vary depending on the dataset.