shawamar / product-recommendation-system-for-e-commerce

using Utility Matrix, TfidfVectorizer, KMeans

product-recommendation-system-for-e-commerce

using Utility Matrix, TfidfVectorizer, KMeans

Notebook with dataset

The recommendation system is designed in 3 parts based on the business context:

Recommendation System - Part I

Product popularity based recommendation system targeted at new customers

• Popularity based are a great strategy to target the new customers with the most popular products sold on a business's website and is very useful to cold start a recommendation engine.
• **Dataset : **Amazon product review dataset

Recommendation System - Part II

Model-based collaborative filtering system

• Recommend items to users based on purchase history and similarity of ratings provided by other users who bought items to that of a particular customer.
• A model based collaborative filtering technique is closen here as it helps in making predictinfg products for a particular user by identifying patterns based on preferences from multiple user data.

Recommendation System - Part III

• For a business without any user-item purchase history, a search engine based recommendation system can be designed for users. The product recommendations can be based on textual clustering analysis given in product description.
• **Dataset : **Home Depot's dataset with product dataset.