Pairwise and Attribute-Aware Decision Tree-Based Preference Elicitation for Cold-Start Recommendation

TL;DR

This paper addresses cold-start recommendations by extending decision-tree based rating elicitation (RE) to gather not only item ratings but also attribute preferences (e.g., genres) and to use item pairs at each decision node. It introduces dynamic tree updating and top-down elicitation, and compares single-item, multi-item-type, and pairwise-item approaches within a music domain. Experiments on Yahoo! Music data show that attribute-aware trees ($tree_hybrid$) improve early recommendations, while pairwise elicitation ($pairwise_tree_2$) yields more informative signals and reduces the number of questions needed. The findings suggest practical pathways to accelerate onboarding of cold-start users and improve CF performance, with future work on optimal pairing, ranking-based learning, online evaluation, and ensemble RE strategies.

Abstract

Recommender systems (RSs) are intelligent filtering methods that suggest items to users based on their inferred preferences, derived from their interaction history on the platform. Collaborative filtering-based RSs rely on users past interactions to generate recommendations. However, when a user is new to the platform, referred to as a cold-start user, there is no historical data available, making it difficult to provide personalized recommendations. To address this, rating elicitation techniques can be used to gather initial ratings or preferences on selected items, helping to build an early understanding of the user's tastes. Rating elicitation approaches are generally categorized into two types: non-personalized and personalized. Decision tree-based rating elicitation is a personalized method that queries users about their preferences at each node of the tree until sufficient information is gathered. In this paper, we propose an extension to the decision tree approach for rating elicitation in the context of music recommendation. Our method: (i) elicits not only item ratings but also preferences on attributes such as genres to better cluster users, and (ii) uses item pairs instead of single items at each node to more effectively learn user preferences. Experimental results demonstrate that both proposed enhancements lead to improved performance, particularly with a reduced number of queries.

Figures (3)

• Figure 1: Illustration of RE procedure
• Figure 2: Decision tree rating elicitation with multiple categories of items
• Figure 3: Decision tree rating elicitation with pairs of items