DoubleShift-AI

Our AI recommendation engine analyzes user behavior,purchases,ratings and wishlists to deliver realtime personalized products by helping customers while boosting engagement and sales.

Comment

Innovhack-SNS

An intelligent AI/ML recommendation engine that delivers personalised product suggestions based on user behaviour and preferences.

🛒 AI-Powered Product Recommendation System

AI-powered e-commerce recommendation system that delivers personalised product suggestions using browsing history, purchases, ratings, and wishlist data. Built with AI/ML for real-time, scalable, and accurate recommendations to improve user experience, engagement, and sales.

Run Locally

Prerequisites: Node.js

  1. Install dependencies: npm install

  2. Set the GEMINI_API_KEY in .env.local to your Gemini API key

  3. Run the app: npm run dev

📖 Table of Contents

About the Project

Problem Statement

Objectives

Features

System Architecture

Technology Stack

AI/ML Model

Dataset

Project Workflow

Installation

Usage

API Endpoints

Folder Structure

Screenshots

Future Scope

Contributors

License

📌 About the Project

Modern e-commerce platforms contain millions of products, making it challenging for customers to discover items that match their interests. Traditional recommendation systems often provide generic suggestions that fail to understand individual customer preferences.

This project introduces an AI-powered recommendation system capable of analysing customer behaviour—including browsing history, purchase history, ratings, and wishlist activity—to generate personalised product recommendations in real time.

The recommendation engine continuously learns from user interactions, enabling businesses to improve customer satisfaction, increase conversions, and maximise sales while maintaining a scalable architecture suitable for large-scale e-commerce platforms.

❓ Problem Statement

Develop an AI-powered recommendation system for an e-commerce platform that provides personalised product suggestions based on customer preferences, browsing history, purchase behaviour, ratings, and wishlist data. The solution should improve product discoverability, customer engagement, shopping experience, and overall business performance while supporting real-time recommendations and scalability.

🎯 Objectives

Improve product discoverability through personalised recommendations.

Enhance customer shopping experience.

Increase customer engagement and retention.

Boost conversion rates and sales.

Reduce product search time.

Learn continuously from customer interactions.

Provide accurate recommendations with low response time.

Build a scalable recommendation engine suitable for enterprise-level e-commerce platforms.

✨ Features

Personalised product recommendations

AI-powered recommendation engine

Collaborative Filtering

Content-Based Filtering

Hybrid Recommendation System

Real-time recommendation generation

Customer behaviour analysis

Wishlist analysis

Purchase history learning

Browsing history analysis

Rating-based recommendations

Trending product recommendations

Similar product suggestions

Scalable architecture

REST API integration

User-friendly dashboard

Fast response time

Continuous model improvement

🏗️ System Architecture

Customer 

                │  

     React Frontend / Web App  

                │  

         REST API (FastAPI)  

                │  

    Recommendation Engine  

                │  

   AI / Machine Learning Model  

                │  

      Feature Engineering  

                │  

    Customer Behaviour Data  

                │  

  PostgreSQL / MongoDB Database  

                │  

  Recommended Products Output

🛠️ Technology Stack

Frontend

React.js

HTML5

CSS3

Tailwind CSS

JavaScript

Backend

Python

FastAPI

Flask (Optional)

Machine Learning

Scikit-learn

Pandas

NumPy

SciPy

Database

PostgreSQL

MongoDB

Deployment

Docker

GitHub

GitHub Actions

Tools

VS Code

Jupyter Notebook

Postman

Figma

🤖 AI/ML Model

The recommendation system combines multiple recommendation techniques to achieve higher accuracy.

Algorithms Used

Collaborative Filtering

Content-Based Filtering

Hybrid Recommendation

Cosine Similarity

K-Nearest Neighbours (KNN)

Matrix Factorisation (Optional)

TF-IDF Vectorisation

Singular Value Decomposition (SVD)

Model Pipeline

  1. Data Collection

  2. Data Cleaning

  3. Feature Engineering

  4. Model Training

  5. Recommendation Generation

  6. User Feedback Collection

  7. Model Retraining

📊 Dataset

The recommendation model uses the following information:

User ID

Product ID

Product Category

Product Description

Ratings

Reviews

Browsing History

Purchase History

Wishlist

Cart Items

Product Popularity

Timestamp

Possible dataset sources:

Kaggle

Amazon Product Dataset

Flipkart Dataset

Custom Dataset

🔄 Project Workflow

Customer Login

Built With

Share this project:

Updates

Submission history