Table 13 Classification accuracy with 95% confidence intervals and McNemar test p-values versus the unified model.

From: A zero-shot LLM framework for multimodal grievance classification, urgency scoring, and abuse detection in civic feedback systems

Model

Accuracy (%)

95% CI

p-value vs unified

Naive Bayes

83.5

[81.1, 85.8]

\(< 0.001\)

SVM

87.1

[85.0, 89.0]

\(< 0.01\)

TF–IDF + rules

76.4

[73.6, 79.1]

\(< 0.001\)

Unified model (Proposed)

92.4

[90.7, 94.0]

–

Back to article page