Clustering Users for Better Content Recommendations

The Naive Approach was the classic k-means algorithm that proved to be the most effective. We created a matrix where each user corresponds to a row and each movie or TV series is represented by a column. The matrix elements were binary, taking the value of 0 or 1 based on whether the user had viewed a particular title.