Movie recommendations with Machine Learning

One of the challenges in recommending new products for a consumer to purchase is that we don't often have a way to deducing what a person tastes are. In order to get a good understanding of what movie recommendations an individual may like, we often have to understand how similar a movie, or a group of movies that were viewed by an individual are, and provide recommendations of these similiarities with the entire gamut of selections in our list.

There are a few ways of providing recommendations of products. One the one hand, we can try to match certain characteristics about items that a consumer bought and then try to compare similar characteristics. Another way is to see if there are similar products bought by two unique users and compare the items that they bought. If the similarity between the purchases of a consumer match within a threshold, then we can suggest an item to one of the consumers that they had not yet bought.

Jupyter Notebook

Attached is a Jupyter Notebook it includes the step by step process of how I broke down the problem of creating a recommendation based on item-based collaborative filtering.

Final thoughts

Let me know what you think! Would this type of item-based collaborative filtering be useful to you? Feel free to contact me through twitter through @paulgarias and continue the conversation here!