CMDPAD: Chinese Multimodal Dynamic Personality and Affect Dataset

CMDPAD challenges the static personality assumption by providing dynamic utterance-level scores for the Big Five personality traits. The dataset moves beyond emotion recognition to predict the emotional trajectory of the next interaction turn. It was authored by HensonXie and last updated on Hugging Face in May 2026.

Use Cases

• Predicting emotional trajectory in conversational turns based on the dataset's predictive focus.
• Modeling dynamic personality traits using utterance-level Big Five scores.
• Training multimodal models with aligned deep features for text and other modalities.

Strengths

• Includes dynamic, utterance-level personality scores, moving beyond static models.
• Focuses on predicting future emotional states, not just recognizing current ones.
• Provides high-quality, aligned multimodal features for text and other modalities.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and overall dataset scale are unknown, limiting suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

HensonXie

Freshness

Last updated 2026-05-14 13:23:10; freshness should be verified.