A Demand Forecasting Model For Women's Sandals in the MSME Supply Chain Using the Linear Regression Algorithm: A Case Study of Ann-D'Mello Sandals

Abstract

Ann-D'Mello Sandals faces challenges in production management due to the lack of systematic production planning and reliance on intuition for demand forecasting. This approach results in inaccurate production quantities and potential losses. This research applies the linear regression method to analyze historical demand data for women's sandals and predict future demand. This method allows for the identification of patterns in historical data, aiding MSME like Ann-D'Mello Sandals in optimizing production based on the relationship between variables such as fashion trends, popularity, seasonality, and other economic factors. The aim of this study is to apply a linear regression algorithm to predict future demand for women's sandals and evaluate the accuracy of these predictions. The results indicate that applying the linear regression algorithm to forecast demand over the next 48 weeks shows an upward trend, with predictions reaching over 4,000 pairs by week 248. This demonstrates a promising market potential for women's sandals and can help MSME in planning more effective production and marketing strategies to meet the increasing demand. The evaluation of the linear regression model shows good performance with an Average MAPE value of 2.79 on the training set and 4.65 on the testing set, using a 10-fold Time Series Cross-Validation (TMCV) scenario. The low MAPE values indicate that the model can predict demand with high accuracy. Overall, this linear regression model has proven effective in forecasting demand for women's sandals, providing valuable guidance for MSME to optimize their production and marketing strategies.