Predictive Analysis of Wholesale Customer Purchases Pattern Using Machine Learning Techniques

Using information from a wholesale distribution company, this study looks at wholesale customer spending trends over several product categories. Emphasizing the various purchasing behaviour among consumers, descriptive statistics and exploratory data analysis expose notable variability and skewness in spending behaviour. Using Linear Regression and Neural Networks, predictive modelling forecasts future consumer purchases; the Neural Network model shows better performance in capturing complex, non-linear relationships. Sobol method sensitivity analysis is done to find the main drivers affecting expenditure over product categories, so improving the predictive accuracy. Strong interdependencies, such the link between Milk and Grocery spending, were underlined by results suggesting chances for focused marketing strategies, product bundling, and better inventory control. The results offer practical knowledge for customer segmenting, best use of resources, and application of customized marketing strategies. This thorough study provides a strong data-driven framework for wholesale distribution decision-making, so improving operational efficiency, reducing waste, and so raising profitability.