Abstract

Wool production industries rely heavily on accurate demand forecasting to manage supply chains and production schedules effectively. Predicting wool demand allows companies to avoid overproduction and resource wastage while meeting market needs efficiently. Traditional forecasting models often struggle with the seasonality and variability of wool demand. The ARIMA (Auto Regressive Integrated Moving Average) model, a time series forecasting technique, is particularly suited for this task due to its ability to capture both trends and seasonal fluctuations in historical data. The data for this research was collected from official records of wool demand across various Indian states for 2016-2023. After data pre-processing (cleaning, sorting, and organizing), the ARIMA model was implemented using Python for analysis and forecasting. Flask was employed to create a web-based interface, allowing users to interact with the predictive model easily. The model parameters (p, d, q) were fine-tuned to fit the data and generate accurate forecasts. The ARIMA model demonstrated a strong ability to predict wool demand, achieving a low Mean Absolute Error (MAE) and Root Mean Square Error (RMSE). The model's forecasts were closely aligned with actual demand figures across different states, capturing both long-term trends and seasonal variations. Graphical representations showed minimal deviation between predicted and actual values, validating the accuracy of the model. The implementation of the ARIMA model for wool demand forecasting has significant practical implications for production planning in the textile industry. Companies can use the system to optimize production schedules, reduce costs, and minimize wastage. This methodology can be extended to other commodities where seasonal demand plays a critical role, offering a versatile tool for various sectors.