Predictive Rate Parity is about?

Predictive Rate Parity is about?

• A. Predictive Rate Parity
• B. Perplexity
• C. Agentic AI
• D. Hallucinations

Answer: (A)

Predictive rate parity focuses on fairness in how often the model flags someone as at risk across different groups. In other words, the rate at which the model produces a positive prediction should be the same for each group, so no group is systematically more or less likely to be warned or referred, just because of group membership. This is about equalizing the frequency of positive predictions, not about equal accuracy or the underlying base risk alone. Achieving this can require different thresholds or adjustments per group, and it can interact with how different groups actually experience risk (base rates), potentially affecting calibration and overall performance. The other items listed aren’t describing this criterion; perplexity relates to language model uncertainty, agentic AI to goal-directed behavior in AI systems, and hallucinations to incorrect outputs.

Predictive rate parity focuses on fairness in how often the model flags someone as at risk across different groups. In other words, the rate at which the model produces a positive prediction should be the same for each group, so no group is systematically more or less likely to be warned or referred, just because of group membership. This is about equalizing the frequency of positive predictions, not about equal accuracy or the underlying base risk alone. Achieving this can require different thresholds or adjustments per group, and it can interact with how different groups actually experience risk (base rates), potentially affecting calibration and overall performance.

The other items listed aren’t describing this criterion; perplexity relates to language model uncertainty, agentic AI to goal-directed behavior in AI systems, and hallucinations to incorrect outputs.