https://github.com/saurabh-takle/personalized-product-recommendation
https://github.com/saurabh-takle/personalized-product-recommendation
Last synced: 2 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/saurabh-takle/personalized-product-recommendation
- Owner: saurabh-takle
- Created: 2025-02-03T17:42:47.000Z (4 months ago)
- Default Branch: main
- Last Pushed: 2025-03-19T20:51:19.000Z (2 months ago)
- Last Synced: 2025-03-19T21:40:30.707Z (2 months ago)
- Language: Jupyter Notebook
- Size: 31 MB
- Stars: 2
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Personalized Product Recommendation System
## **Project Overview**
This project implements a **Personalized Product Recommendation System** using **collaborative filtering**. The goal is to provide product recommendations based on user interactions, leveraging machine learning techniques. The model is trained on a dataset containing user-product interactions and is saved for future use.
## **Features**
- Data preprocessing and exploratory data analysis (EDA)
- Model training using **collaborative filtering** with Surprise library
- Evaluation using RMSE (Root Mean Squared Error)
- Saving and loading trained models
- Generating personalized recommendations
---
## **Project Structure**
```
📂Personalized-Product-Recommendation/
├── 📂data/ # Datasets used for training/testing
├── 📂notebooks/ # Jupyter Notebooks for EDA & training
├── 📂models/ # Saved trained models
├── 📂src/ # Core Python scripts
│ ├── data_processing.py # Data loading & preprocessing
│ ├── train_model.py # Training collaborative filtering model
├── README.md # Project documentation
├── requirements.txt # Dependencies
```
---
## **Dataset**
The dataset contains user interactions with products, including ratings or implicit feedback. We process and split the data into training and testing sets before applying collaborative filtering techniques.
---
## **Tech Stack**
- **Python**
- **Jupyter Notebook**
- **Pandas, NumPy** (Data Handling)
- **Surprise** (Collaborative Filtering)
- **Matplotlib, Seaborn** (Data Visualization)
- **Pickle** (Model Saving & Loading)
---
## **Setup Instructions**
### 1. Clone the Repository
```bash
git clone https://github.com/saura-t/personalized-product-recommendation.git
cd personalized-product-recommendation
```
### 2. Create a Virtual Environment (Recommended)
```bash
conda create --name [venv_name] python=3.9 -y
conda activate [venv_name]
```
### 3. Install Dependencies
```bash
pip install -r requirements.txt
```
---
## Data Preprocessing
### Load the Dataset
```python
import pandas as pd
df = pd.read_csv('data/reviews.csv')
print(df.head())
```
### Check for Missing Values
```python
print(df.isnull().sum())
df.dropna(inplace=True)
```
### Data Visualization
```python
import matplotlib.pyplot as plt
import seaborn as sns
plt.figure(figsize=(10,5))
sns.countplot(x='interaction_type', data=df)
plt.title('Distribution of Interaction Types')
plt.show()
```
---
## Model Training
### Train-Test Split
```python
from surprise import Dataset, Reader
from surprise.model_selection import train_test_split
reader = Reader(rating_scale=(1, 5))
dataset = Dataset.load_from_df(df[['user_id', 'product_id', 'rating']], reader)
trainset, testset = train_test_split(dataset, test_size=0.2)
```
### Train Model
```python
from surprise import SVD
from surprise import accuracy
model = SVD()
model.fit(trainset)
```
### Evaluate Model
```python
predictions = model.test(testset)
rmse = accuracy.rmse(predictions)
print(f'RMSE: {rmse}')
```
---
## Model Training with Hyperparameter Tuning
We perform hyperparameter tuning using **GridSearchCV** to find the best parameters for the SVD model.
```python
from surprise.model_selection import GridSearchCV
from surprise import SVD
param_grid = {
'n_factors': [50, 100, 150],
'n_epochs': [10, 20],
'lr_all': [0.002, 0.005],
'reg_all': [0.02, 0.1]
}
grid_search = GridSearchCV(SVD, param_grid, measures=['rmse'], cv=3)
grid_search.fit(data)
# Best Parameters
print("Best RMSE Score:", grid_search.best_score['rmse'])
print("Best Parameters:", grid_search.best_params['rmse'])
```
---
## Save & Load Model
### Save Model
```python
import pickle
with open('models/recommendation_model.pkl', 'wb') as f:
pickle.dump(model, f)
```
### Load Model
```python
with open('models/recommendation_model.pkl', 'rb') as f:
loaded_model = pickle.load(f)
```
---
## Model Evaluation
- Performance is measured using **Root Mean Squared Error (RMSE)**.
- A good RMSE value depends on the dataset but should ideally be **below 1.0**.
```python
from surprise import accuracy
predictions = best_model.test(testset)
rmse = accuracy.rmse(predictions)
print("Final RMSE:", rmse)
```
---
## Author
Your Name
[LinkedIn Profile](https://linkedin.com/in/saurabhtakle)
[GitHub](https://github.com/saura-t)