Abstract
Session-based recommendations aim to predict a user’s next click based on the user’s current and historical sessions, which can be applied to shopping websites and APPs. Existing session-based recommendation methods cannot accurately capture the complex transitions between items. In addition, some approaches compress sessions into a fixed representation vector without taking into account the user’s interest preferences at the current moment, thus limiting the accuracy of recommendations. Considering the diversity of items and users’ interests, a personalized interest attention graph neural network (PIA-GNN) is proposed for session-based recommendation. This approach utilizes personalized graph convolutional networks (PGNN) to capture complex transitions between items, invoking an interest-aware mechanism to activate users’ interest in different items adaptively. In addition, a self-attention layer is used to capture long-term dependencies between items when capturing users’ long-term preferences. In this paper, the cross-entropy loss is used as the objective function to train our model. We conduct rich experiments on two real datasets, and the results show that PIA-GNN outperforms existing personalized session-aware recommendation methods.
Highlights
Entropy 2021, 23, 1500. https://In recent years, the rapid development of Internet technology in many applications has caused information overload
We use the personalized graph neural networks (PGNN) to learn complex transformational relationships between items that have interacted with the user, and obtain the representations of items and users
PIAGNN incorporates user embeddings to improve the personalized representation of the model, uses goal-aware attention to explore user interest preferences to enrich the graph-based model, and activates different user interests for different target items to improve the expressiveness of the recommendation model
Summary
The rapid development of Internet technology in many applications (such as e-commerce, social media, etc.) has caused information overload. In actual applications, user information is unknown, and user history behaviors are not available besides the current session, so it is difficult to produce accurate results To solve these problems, a session-based recommendation system extracts interactive information to represent the user’s preference transfer and use the limited historical behavior to predict the user’s action (such as which item to click). Complex user behavior patterns are important for session-based recommendation, yet the above sequence-based approach only modeled sequential transitions between consecutive items, ignoring the relationship between other items To overcome this limitation, this paper models the items in a session as a session graph, captures the transitions between items by GNN, and generates an embedding vector representation. We use a new user-based personalized graph neural network (PGNN), which can capture complex item transformations for different user interests.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
