Matthews correlation coefficient (MCC) measures balanced metric even with class imbalance
Image: NormanEinstein, CC BY-SA 3.0, via Wikimedia Commons
Phi coefficient
Matthews correlation coefficient (MCC) measures balanced metric even with class imbalance
The Matthews correlation coefficient (MCC) is a balanced metric that accounts for true and false positives and negatives, making it effective even when classes are imbalanced.
MCC considers all four combinations of the predicted and actual binary classifications: true positives, true negatives, false positives, and false negatives. This comprehensive approach ensures that the performance metric reflects the actual effectiveness of a binary classification model.
Unlike other metrics that may favor the majority class, MCC provides a balanced view of model performance across both classes. This makes it particularly useful for evaluating models in scenarios where class imbalance is present.
Example
In a binary classification problem with 100 positive cases and 900 negative cases, a model with an MCC of 0.7 indicates good balanced performance across both classes, regardless of the imbalance.
Remember this
Understanding MCC helps in choosing the right metric for evaluating model performance in imbalanced datasets, ensuring accurate and fair assessment.
Text adapted from Wikipedia, licensed under CC BY-SA 4.0.
Related concepts
Pearson correlation coefficient
Pearson correlation coefficient formula: r = Σ[(xi - x̄)(yi - ȳ)] / [√(Σ(xi - x̄)²) * √(Σ(yi - ȳ)²)]
to use F1 score: when classes are imbalanced and both FP and FN matter
Why might a sports team need a different score to judge their performance?
Markov chain Monte Carlo
MCMC samples from complex posterior distributions
Top-k vs top-p sampling: top-k fixes candidate count, top-p fixes cumulative probability mass
Top-k sampling fixes candidate count; top-p sampling fixes cumulative probability mass
the dot product measures alignment: it equals |a||b|cos(θ)
Why do vectors sometimes "agree" with each other?
to use AUC-ROC: comparing classifiers across all thresholds
Why can't we always trust a yes-or-no answer?
Swipe through 100 ML concepts daily
Open Pocket Polymath