pranshu1921 / amazon-fashion-discovery-engine

Recommend similar apparel searched by user on Amazon.

Amazon-Fashion-Discovery-Engine

This is a content based recommendation engine for recommending apparel items or products at Amazon, using text and image data retreived from website. Suggested text based recommendations using Bag of Words (BoW), Word2Vec and TF-IDF techniques. Made image based recommendations using Convolutional Neural Network(CNN).

Table of contents

• General info
• Setup
• Status
• Inspiration
• Contact

General info

Amazon, each year makes additional 30% revenue through product recommendations, a value greater than $40 Billion.

Using this content based recomendation, we get a glimpse to the recommendation systems at Amazon, constantly recommending the right products to customers.

We made use of text and image data of products, web scraped for the website using the Amazon API.

Brand name, title, description, price are one of the major features for extraction of data for recommendation engine.

Setup

• Download the Jupyter Notebooks on your computer.
• Download training, testing, and other important data using link provided in this file.

Install the requirements

• Install the requirements using pip install -r requirements.txt.

  • Make sure you use Python 3.

• Run the jupyter notebooks in the following order -

  1. AppliedAIWorkshop.ipynb
  2. image_similarity_cnn.ipynb

Status

Project status: Finished

Inspiration

This project was part of one of the Applied Machine Learning Case Studies provided by AppliedAIcourse.

Contact

Feel free to contact me, send a mail to [email protected]