Personalized Fashion Prediction Using Multimodal Fusion and Collaborative Filtering Algorithms

In recent years, with advancements in information technology and material standards, the amount of data related to clothing products has increased significantly. However, fashion consumers often struggle to quickly identify products that meet their needs due to the overwhelming volume of available data. Traditional recommendation methods typically focus on a single attribute and fail to capture the complex features of clothing and the diverse needs of consumers. Multi-modal fusion technology, by integrating various data sources such as images, text, allows for a deeper exploration of clothing's multidimensional characteristics. Collaborative filtering algorithms, on the other hand, offer personalized recommendations based on similarities between users or items. Despite significant progress in both areas, combining these technologies with personalized clothing recommendations remains in the early stages. This paper reviews the latest research, highlighting challenges and proposing innovative solutions, with the aim of improving the accuracy and efficiency of personalized clothing recommendation systems to meet consumers' evolving demands.