calculate_metrics: Calculate Association Metrics

View source: R/calc_assoc_metrics.R

calculate_metricsR Documentation

Calculate Association Metrics

Description

Helper function that computes various association metrics for bigrams based on their probability distributions. Supports PMI (Pointwise Mutual Information), Dice's Coefficient, and G-score calculations.

Usage

calculate_metrics(bigram_probs, association)

Arguments

bigram_probs

A data frame containing bigram probability data with columns:

  • p_xy Joint probability of bigram

  • p_x Marginal probability of first token

  • p_y Marginal probability of second token

association

Character vector specifying which metrics to calculate

Value

A data frame containing the original probability columns plus requested association metrics:

  • pmi: Pointwise Mutual Information

  • dice_coeff: Dice's Coefficient

  • g_score: G-score


qtkit documentation built on April 4, 2025, 4:47 a.m.