👍Style Recommendation

Another popular area where AI is used for Fashion World is style recommendation. Since the significant data era, choosing what to wear has become complicated and timeconsuming because of many choices. Thanks to advances in AI, the selection is made by AI, and AI can scan many items and choose the best matching items for us.

Recommender system study is a highly active research field, and recommenders are already used in many online shops, including Amazon, Google Shopping, and Shop It To Me. There is research for location-based style recommendations [17]. Most of these researches focus usage of CNN and derived models for advice [16].

Another research area related to style recommendation is fashion compatibility, and these systems predict whether different fashion items go together or not. Several tasks are at hand for this research area: recommending all clothing items given one thing or recommending missing clothing provided multiple clothing items. For this application, again, CNN and derived models are mostly preferred. However, models like LSTM, which is used for sequential data, are also used with accuracy reaching up to 90% percentage [18,19].

Another research area of style recommendation is personal style recommendation. Personel Style Recommendation Models learn from unique style preferences and recommend new clothing or items for a set of items. This area is trendy because of its benefits fortis usage in shopping. CNN and derived models are the most popularly used models for this application, and their success rate goes up to 80% accuracy [20].

Fashionic AI aims to recommend style and clothing combinations according to occasion and environment, using its users' personal choices. Fashionic AI brings a novel style recommendation by combining these two separate recommendation systems.

Clustering in Fashionic AI