Time Series Forecasting and Modeling of Food Demand Supply Chain Based on Regressors Analysis

Aim:

       To Develop a methodology that combines the robustness of ARIMA and SARIMA models with the explanatory power of regressors analysis to improve forecasting accuracy within the food supply chain

Abstract:

       This study aims to improve the accuracy of forecasting and modeling in the food demand-supply chain by integrating ARIMA (AutoRegressive Integrated Moving Average) and SARIMA (Seasonal ARIMA) models with regressors analysis. With the growing complexity of the food supply chain and the increasing demand for accurate forecasts, traditional time series models often fall short in capturing the intricate dynamics influenced by various external factors. In this research, we propose a novel approach that combines the robustness of ARIMA and SARIMA models with the explanatory power of regressors analysis. By incorporating relevant external variables such as economic indicators, weather patterns, and demographic factors, the model aims to better capture the underlying drivers of food demand and supply.The methodology involves several steps: data collection and preprocessing, exploratory data analysis to identify potential regressors, model fitting using ARIMA and SARIMA with regressors, validation using appropriate metrics such as Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE), and finally, forecasting future demand-supply dynamics

 Existing Method:

      The ‘Food Demand Forecasting’ dataset released by Genpact, compares the effect of various factors on demand, extracts the characteristic features with possible influence, and proposes a comparative study of seven regressors toforecast the number of orders. In this study, we used Random Forest Regressor, Gradient Boosting Regressor (GBR), Light Gradient Boosting Machine Regressor (LightGBM), Extreme Gradient Boosting Regressor (XGBoost), Cat Boost Regressor, Long Short-Term Memory (LSTM), Bidirectional LSTM (BiLSTM) inparticular. The results demonstrate the potential of deep learning models in forecasting and highlight the superiority of LSTM over other algorithms. The Root Mean Squared Log Error( RMSLE), Root Mean Square Error (RMSE),Mean Average Percentage Error( MAPE), and Mean Average Error(MAE)

Proposed Method:

         This proposed system aims to leverage the Dickey-Fuller test for stationarity analysis and subsequently employ ARIMA (Auto Regressive Integrated Moving Average) and SARIMA (Seasonal ARIMA) models for time series forecasting in food supply chains. The system integrates these techniques to enhance forecasting accuracy, enabling stakeholders to make informed decisions regarding inventory management, resource allocation, and planning