Optimizing Fashion Recommendation Systems: A Study of Deep Learning Models and Distance Metrics

Deep learning algorithms are increasing in popularity for prediction tasks, indicating varying degrees of accuracy. Deep learning, a branch of artificial intelligence (AI), empowers systems to learn and enhance their performance through experience, without the need for explicit programming. This study examines and compares the performance of different deep learning algorithms in terms of fashion recommendation systems. Specifically, we evaluate the effectiveness of using the cosine and Euclidean distances mathematically combined with deep learning models ResNet50 and VGG16. Furthermore, we discuss the results by showing the original images, suggested items, and the distances between them, providing useful information about the system’s functionality and the reasons behind each recommendation.