From Affect Prediction to Affect Forecasting: Evidence for Distinct Information Sources in Longitudinal Text
2026-06-15 • Artificial Intelligence
Artificial IntelligenceComputation and Language
AI summaryⓘ
The authors studied how to predict current feelings (valence and arousal) from people's written journal entries and how to forecast changes in those feelings over time. They found that understanding the meaning in text helps estimate current feelings fairly well. However, predicting how feelings will change next is better done using simple past numeric data about previous feelings rather than the text itself. Their new models perform well for current feeling prediction but less so for forecasting future changes from text alone. This suggests that texts and numeric mood histories provide different useful information for these two tasks.
valencearousalaffect predictionaffective change forecastinglongitudinal text analysisTrait-State Affective Prediction (TSAP)ecological momentary assessmentPearson correlationtextual semanticsnumeric trajectory features
Authors
Sadia Noor, Seemab Latif, Raja Khurram Shahzad, Mehwish Fatima
Abstract
Modeling dimensional affect in longitudinal text requires distinguishing current affect estimation from future affective change forecasting. Existing approaches often treat each text as an independent observation and apply similar assumptions to both tasks, without testing whether they rely on different information sources. This paper investigates that distinction using longitudinal self-reported ecological essays and feeling-word entries. We propose the Trait--State Affective Prediction (TSAP) framework and its temporal extension E-TSAP for per-text valence and arousal prediction, evaluated on a held-out prediction test set of 1,737 entries from 91 users. We further propose the Affective Change Forecaster Hybrid (ACF-Hybrid) for next-step affective change forecasting, evaluated on a held-out forecasting test set of 46 users. For prediction, E-TSAP achieves composite Pearson correlations of 0.670 for valence and 0.449 for arousal. For forecasting, textual representations perform worse than compact numeric trajectory baselines: the text-inclusive model achieves only r=0.316 for valence and r=0.284 for arousal, whereas a simple prior-state baseline reaches r=0.615 and r=0.670, respectively. ACF-Hybrid, using dimension-specific numeric trajectory features, achieves r=0.659 for valence and $r=0.658$ for arousal. These results show that textual semantics support current affect prediction, whereas future affective change is better captured through prior numeric trajectory dynamics.