TRUST: Item-Calibrated Interval Evidence for Temporal Session-Based Recommendation

2026-06-25

AI Summary

This work addresses a key limitation in existing session-based recommendation methods, which typically assume uniform interest semantics across all items for identical time intervals, thereby ignoring the distinct temporal interval distributions inherent to different items. To overcome this, the authors propose TRUST, a novel framework that introduces, for the first time, an item-level time interval calibration mechanism. This mechanism standardizes observed intervals relative to each item’s empirical distribution, yielding a model-agnostic calibration scoring function that can be flexibly integrated into neighbor sampling, session graph encoding, and interest aggregation modules. Extensive experiments demonstrate that TRUST significantly outperforms state-of-the-art temporal and non-temporal baselines across multiple public datasets. Plug-in validation confirms the general effectiveness of the proposed scoring function, while ablation studies further verify that temporal calibration consistently enhances performance across all incorporated modules.

Abstract

Temporal signals have been widely used in session-based recommendation to infer user interest. Existing temporal session-based recommenders primarily rely on absolute interval values, implicitly assuming that the same interval carries similar interest signals across items. However, we empirically find that this assumption does not hold: each item has its own interval distribution, so an interval should be interpreted relative to the item it belongs to. Based on this observation, we propose TRUST, a framework that evaluates each observed interval relative to the empirical interval distribution of the corresponding item. Specifically, we propose a score function to guide global neighbor sampling, session graph encoding, and final interest aggregation. Experiments on public datasets show that TRUST consistently improves over representative temporal and non-temporal baselines, and plug-in experiments further show that the proposed scoring function can improve existing temporal session recommenders as a model-agnostic method. Component-wise ablations further show that calibrating the temporal signals within each module, rather than removing the module itself, consistently improves neighbor sampling, session graph encoding, and interest aggregation.

Problem

• temporal session-based recommendation
• time interval
• item-specific distribution
• user interest modeling
• session recommendation

Innovation

• temporal session-based recommendation
• interval calibration
• item-specific distribution
• model-agnostic enhancement
• session graph encoding