PREDICTING DEMAND FOR RETAIL PRODUCTS OF INDORAMA SYNTHETICS LTD, BUTIBORI, NAGPUR

Abstract

In the dynamic and competitive textile industry, understanding and anticipating consumer demand is critical for maintaining operational efficiency and profitability. This research explores demand prediction for retail products at Indorama Synthetics Ltd., located in Butibori, Nagpur. By leveraging data-driven techniques, including time-series forecasting and machine learning models, the study aims to provide actionable insights that can help the organization optimize inventory, streamline production schedules, and enhance customer satisfaction. The objective is to identify which predictive model offers the highest accuracy in forecasting demand patterns for key retail products. The study considers various demand-influencing factors such as seasonal trends, economic conditions, product category, and regional sales data.

The methodology involves collecting secondary data from company records, retail sales reports, and relevant industry publications. Data analysis includes the application of statistical tools such as ARIMA, linear regression, and machine learning algorithms like Random Forest and XGBoost. The models are trained and validated using historical sales data, and their performance is evaluated using metrics such as Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and R² Score.

Findings indicate that while traditional statistical models perform reasonably well, ensemble machine learning models yield significantly higher prediction accuracy due to their ability to capture nonlinear relationships and handle multivariate data efficiently. Among the tested models, XGBoost showed the best performance with the lowest prediction error and highest reliability.

The implications of this research are multi-faceted. Firstly, it enables Indorama Synthetics Ltd. to make informed decisions regarding stock replenishment and procurement planning. Secondly, it aids in demand planning during promotional campaigns and seasonal peaks. Thirdly, the study contributes to academic literature by demonstrating a practical implementation of AI/ML models in textile demand forecasting within a regional Indian industrial context.

In conclusion, predictive demand modeling stands out as a strategic tool for supply chain and inventory management. The study recommends further development of integrated demand forecasting systems supported by real-time data analytics. Future research may explore hybrid forecasting frameworks and real-time integration with ERP systems to further enhance predictive capabilities.