Intelligent Sales Prediction System Using Machine Learning

The Intelligent Sales Prediction System is a machine-learning-based application designed to estimate future sales by analysing historical sales data. Sales forecasting supports business activities such as inventory planning, production scheduling, resource allocation, budgeting and setting realistic sales targets. In many small and medium-sized businesses, sales estimates are prepared through manual calculations or basic trend observations. Such methods may not identify complex relationships among product demand, price, quantity, seasonality, discounts and other influencing factors. The proposed system applies data preprocessing and regression-based machine-learning techniques to discover patterns in the available dataset and generate sales predictions. The workflow includes dataset collection, data inspection, cleaning, feature preparation, training and testing data separation, model development, prediction and performance evaluation. Linear Regression is used to estimate the relationship between selected input variables and the sales target, while Decision Tree Regression learns decision rules that can represent non-linear patterns. The models are evaluated using Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE) and the coefficient of determination (R²). The system presents predicted sales and evaluation results in a clear form to support data-driven decisions. The prediction is an estimate based on historical patterns and does not guarantee exact future sales. Its usefulness depends on the quality, quantity and relevance of the dataset, selected features and evaluation method. The proposed approach can be extended with additional algorithms, updated data and interactive visualizations.