Bhuvi18/business-sales-forecast-models

Business Sales & Revenue Forecasting — ML Models

This repository contains the trained machine learning model artifacts used by the Business Sales & Revenue Forecasting System.

The models are used by the Flask backend to generate weekly sales forecasts for stores and departments.

Project

Business Sales & Revenue Forecasting System

The complete application combines:

• React frontend
• Flask backend
• Machine learning forecasting
• Cold-start forecasting
• Store and department master data
• Analytics
• Forecast history

The application supports forecasting for both established and newly introduced business entities.

---

Model Files

File	Purpose
xgboost_model.joblib	XGBoost regression model used for sales forecasting
random_forest_model.joblib	Random Forest regression model
linear_regression_model.joblib	Linear Regression baseline model
ann_model.joblib	Artificial Neural Network regression model
ann_scaler.joblib	StandardScaler used to preprocess ANN inputs

---

Forecasting Approach

The system uses different strategies depending on data availability.

1. Historical ML Forecasting

For an existing store and department with sufficient historical sales:

Historical Sales
       ↓
Feature Engineering
       ↓
Lag Features
Rolling Features
Calendar Features
Store/Department Features
       ↓
Machine Learning Model
       ↓
Weekly Sales Forecast