Artificial Intelligence (AI)-Driven Business Intelligence for Enhancing Retail Performance with Customer Insights

Abstract

Abstract—The primary goal of a retail company's business intelligence (BI) deployment is to examine the advantages of BI adoption as well as the challenges that arise.  The research is on a big-box retailer that sells a variety of goods and has operations in several nations. Online fashion retail has seen rapid growth, necessitating efficient and accurate clothing classification for improved customer experience. This research introduces a method for classifying garments using a tagged dataset of purchase data that relies on Convolutional Neural Networks (CNNs).   A CNN model is trained using a SoftMax layer for classification after convolutional and pooling layers for feature extraction.  The dataset is divided 80:20 between testing and training. In terms of accuracy (94.6%), precision (96.8%), recall (96.8%) and F1-score performance, the CNN model performs better than benchmark models such as InceptionV3.  The suggested model's ability to increase classification accuracy for online fashion retail applications is validated by the comparison study. Through improved product categorization and user experience, this research demonstrates the potential of deep learning models to improve the classification of clothes sold online.