Outfit Recommendation System Using Visual Compatibility and Co-Purchase Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern online retailers face challenges in recommending complementary items for outfits, particularly in the apparel space, as existing recommendation systems struggle to effectively suggest items that match the style and functionality of an anchor item.
Innovation Solution
A system and method for determining complementary items for outfit recommendations, which involves selecting look templates, generating preliminary looks by selecting non-accessory items, matching accessory items using machine learning, and refining recommendations based on inventory and user feedback, while employing graph-based similarity and co-purchase signals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing recommendation systems are used to suggest complementary items, then the system can provide basic item recommendations, but the recommendations fail to effectively match the style and functionality of the anchor item
Solution Approach 1:
The recommendation system is segmented into multiple specialized modules: graph-based similarity module for visual/style matching, co-purchase signal module for functional compatibility, and machine learning module for accessory recommendations. Each module handles specific aspects of complementarity, improving overall recommendation accuracy while maintaining style and functionality adaptability.
Solution Approach 2:
The system changes the parameters used for recommendation by incorporating visual embeddings, graph-based similarity metrics, and co-purchase signals alongside traditional collaborative filtering. This multi-parameter approach enables both accurate matching and adaptability to different styles and functionalities.
2Productivity
If a comprehensive recommendation system is built to provide personalized outfit recommendations, then user engagement and purchase rates improve, but system complexity increases
Solution Approach 1:
The system merges multiple recommendation approaches (graph-based similarity, co-purchase signals, machine learning) into a unified framework that processes anchor items and generates complementary recommendations through integrated modules, achieving high user engagement while managing complexity through modular design.
Solution Approach 2:
The system introduces intermediate processing layers including image embedding generation, graph-based similarity computation, and machine learning model inference as mediators between the anchor item and final recommendations. These intermediaries break down the complex task into manageable steps, improving productivity while controlling system complexity.
Data Source
AI summary
A computer-implemented method including determining, based on an anchor item, at least one look template from a plurality of look templates. The at least one look template can include an anchor super product type for the anchor item, one or more remaining non-accessory super product types, and one or more accessory super product types. The method also can include determining one or more respective complementary items for the anchor item in each of the one or more remaining non-accessory super product types to generate one or more preliminary looks. The method additionally can include determining, via a machine learning module, at least one respective accessory recommendation for the anchor item for each of the one or more preliminary looks based at least in part on respective visual compatibility of the at least one respective accessory recommendation with respective existing items of each of the one or more preliminary looks to create one or more looks. The method further can include transmitting, via a computer network, the one or more looks to be displayed on a user interface for a user. Other embodiments are described.