aksace/superkart-sales-forecasting-model

SuperKart Sales Forecasting Model

This is a Random Forest Regressor model trained to predict product sales at SuperKart stores.

Model Details

• Model Type: Random Forest Regressor
• Task: Regression (Sales Forecasting)
• Training Data: SuperKart historical sales data (8,763 records)
• Test R² Score: 0.9319
• Test RMSE: $278.68

Best Hyperparameters

• n_estimators: 200
• max_depth: None
• min_samples_split: 5
• min_samples_leaf: 2

Features

The model uses the following features:

• Product_Weight
• Product_Sugar_Content
• Product_Allocated_Area
• Product_Type
• Product_MRP
• Store_Size
• Store_Location_City_Type
• Store_Type
• Store_Age
• Price_Category

Usage

import joblib
import pandas as pd

# Load model
model = joblib.load('best_model.pkl')

# Load label encoders
label_encoders = joblib.load('label_encoders.pkl')

# Make predictions
predictions = model.predict(X_test)

Performance Comparison

Model	Test R² Score	Test RMSE
Random Forest	0.9319	$278.68
XGBoost	0.9314	$279.69
Gradient Boosting	0.9290	$284.58

Training Details

• Train-Test Split: 80-20
• Cross-Validation: 3-fold
• Evaluation Metrics: RMSE, MAE, R²