Building on TuriCreate model creation.
Turicreate is a really powerful Python library that can help in building seamless Machine Learning models in the fields of object detection, recommender system, image classification etc.
In this repository I have TuriCreate library to develope a Basic Movie Recommender System. Recommender Systems can be broadly of 2 types:
1.Content Based Recommender Systems: These type of recommender Systems suggest products based on the input data of the user, thus spanny to a large user base. Itanswers the principle question of "What are the similar products?".
2.Collaborative Filtering Recommender Systems: These type of Recommender Systems answer the question, "Who else?". This means that the model will recommend the products (in this case movie names) that are similar to the other users of the service. It works on the basic principle of users buying similar products today will prefer buying more such common products. They are mostly used in social, consumer-friendly enviornments.Thy are the more popular form of Recommender Systems.
One of the main problems that Recommender Systems face is called Cold Start problem. When a new user, signs up for the service, what recommendations could the service provide him/her? Or the more important question that is, "How do we recommend our products to her/him? and which products?"
This problem is overcome Knowledge-Based Recommender Systems that capture digitally stored knowledge in a company’s backend to match specific user queries by decoding its intents, context, and entities in their backend in the form of datasets. They then apply this to a new user on the basis of an "if-this-then-that" approach.Once the user begins to use the services of the company, the Cold Start problem slowly diminishes.
The following model uses Content Based Recommender Systems.