This project is from one of the case studies of AppliedAI Course (https://www.appliedaicourse.com/course/13/workshop) . The commited changes in the ipynb build up a solution to get apparel recommendations from Amazon's products. The data from Amazon.com is used for learning purpose to get best solution for proposed problem.
To summarize, the product similarities are calculated by using euclidean distance, cosine similarity, symantic similarities(w2v) and cnn(vgg16) feature extraction with product's title, color and brand name as features. The ipynb is very much self explanatory.