EAC-Agent Model Performance Metrics on IEMOCAP and MELD Datasets

76.27% accuracy on IEMOCAP and 67.57% on MELD for emotion recognition. The dataset contains performance metrics for a multimodal conversational agent (EAC-Agent) compared against existing techniques on two benchmark datasets. It was authored by Shahid Jamil and last updated on April 17, 2026.

Use Cases

• Benchmarking multimodal emotion recognition models based on reported accuracy scores
• Evaluating emotion-aware response generation based on perplexity, BLEU, and ROUGE-L metrics
• Comparing transformer-based sequence-to-sequence models with pre-trained embeddings like GloVe
• Assessing the impact of self and cross-modal attention on text, audio, and visual features

Strengths

• Provides specific performance metrics: 76.27% accuracy on IEMOCAP and 67.57% on MELD
• Includes multiple evaluation scores for response generation: perplexity (39.01, 42.30), BLEU (0.31, 0.30), ROUGE-L (0.45, 0.44)
• Results are validated against two established benchmark datasets (IEMOCAP and MELD)

Limitations

• Column-level documentation is absent; field semantics must be inferred after download
• Row count is unknown, which may limit suitability assessment
• The dataset is 5.5 KB, indicating a very limited scope likely containing only summary metrics

Provenance

Source

figshare

Collection Method

Likely contains aggregated performance metrics from model validation experiments.

Freshness

Last updated 2026-04-17 17:35:26; freshness should be verified