Exécution de notre outil de synthèse sur un nouvel article

Résumé : The fashion retail industry faces challenges in demand forecasting, inventory management, and pricing strategies due to the complexity of consumer behaviour and market dynamics. This study employs advanced data mining techniques, including Linear Regression, Decision Tree Regressor, Random Forest, Gradient Boosting, and Support Vector Regression, to analyse sales data and develop predictive models. Comprehensive preprocessing addressed missing values, outliers, and data inconsistencies to ensure reliability. Exploratory Data Analysis (EDA) revealed key insights into purchase behaviour, item popularity, and seasonal trends, identifying factors influencing sales performance. Models were evaluated using metrics like RMSE, MAE, and R-squared, with Random Forest emerging as the most effective for balancing accuracy and generalizability. Findings highlighted actionable patterns, such as the impact of holidays on sales spikes and the role of payment preferences in consumer behaviour. These insights offer solutions for optimizing inventory management, improving efficiency, and reducing waste. This research supports sustainable practices in the fashion retail sector and aligns predictive analytics with real-world applications. Future work will integrate multi-modal data sources, like social media trends, and explore adaptive models for real-time decision-making.