Explainable Prediction from Mobile Sensing Data through LLM-guided Concept Integration

A Concept-Integrated Transformer (CIT) using LLM-guided concept supervision achieved the highest F1 score of 0.756 on the AFFECT dataset and tied for the highest at 0.765 on a PHQ-9 dataset, according to a new arXiv paper (2609.11995v1). The method uses a pretrained large language model to generate baseline-aware concept abnormality targets with confidence weights, avoiding manual concept annotation. In AFFECT, sleep quantity and quality showed the clearest difference between high and low negative affect groups, supporting LLM-guided concept integration for interpretable prediction in small-cohort mobile sensing studies.

arXiv:2609.11995v1 Announce Type: new Abstract: Mobile sensing enables longitudinal monitoring of behavioral and physiological patterns in everyday settings. However, accurate prediction remains challenging in small-cohort health-sensing studies, where task-specific outcome supervision is limited relative to heterogeneous sensing data. Interpretability is also important, as model outputs should reflect meaningful behavioral and physiological patterns rather than predictive scores alone. We develop a Concept-Integrated Transformer CIT with LLM-guided concept supervision for explainable prediction from mobile sensing data. CIT uses a pretrained large language model to generate baseline-aware concept abnormality targets with confidence weights without manual concept annotation. Across two longitudinal datasets, CIT achieves the highest F1 score on AFFECT 0.756 and ties for the highest on a PHQ-9 dataset 0.765 . The learned concept scores also reveal interpretable behavioral and physiological patterns; in AFFECT, sleep quantity and quality show the clearest difference between high and low negative affect groups. These findings support LLM-guided concept integration for accurate and interpretable prediction in small-cohort mobile sensing studies.