Demand Forecasting

This project is used to learn about different time series forecasting techniques for store-item product forecasting using Kaggle Dataset.

-- Project Status: [Active]

Description

This module is designed to go over different techniques used in Time-Series Analysis and Forecasting.

Methods Implemented

ARIMA Model

AutoRegressive Integrated Moving Average (ARIMA) model. Implemented using the statsmodels library in Python.

Facebook's Prophet

Facebook's Prophet is an open-source library used for time series forecasting. Facebook's Prophet is available as a Python package.

Exponential Smoothing

Implement Simple Exponential Smoothing, Double Exponential Smoothing, and Triple Exponential Smoothing.

Machine Learning Models

Trained a LightGBM Regressor model and performed feature engineering to improve accuracy of model. Feature engineering techniques include: Lage Features, Rolling and Expanding Window Features, Date-Time Features).