A basic approch from NLP and Computer vision to get list of similar products on clothes dataset from amazon api. Text and image analysis are performed to get high level features.
Text analysis on title, brand and color with the help of following model
- Baggage of words
- tf-idf
- word2vec
- word2vec*idf
- data from last flatten layer of VGG16 is considered because it was able to recognize pattern/small object from clothes.
On the basis of euclidean distance between numerical features between queried product and remaining products, similar products are found out.
First one is query image and rest are recommendations
