Distributional sentiment modeling and anomaly detection for consumer complaint assessment
arXiv:2609.31653v1 Announce Type: new Abstract: Sentiment analysis is a common tool for converting unstructured text into quantitative signals in finance and risk management. Yet most applications reduce the output to a discrete polarity label or a single predictive feature, overlooking the distrib…
Read the full story at arXiv cs.CL ↗
Timeline · 1 report
- 2026-09-29 04:00 · arXiv cs.CL
Distributional sentiment modeling and anomaly detection for consumer complaint assessment