An ML team is analyzing the breakdown of a classification model's errors. What is the name of the tool used to organize prediction results into a table of the combinations of actual class and predicted class (true positive, false positive, true negative, false negative) so that error tendencies can be understood?
Choosing the table that organizes prediction results (the confusion matrix).
ROC curve
An ROC curve is a curve that plots the relationship between the true positive rate and the false positive rate as the threshold is varied (true positive rate = the proportion of actual positives correctly judged as positive, false positive rate = the proportion of actual negatives incorrectly judged as positive).
It is a graph, not a table that tabulates the four combinations (the confusion matrix), so it is incorrect.
Scatter plot
A scatter plot is a general visualization that plots the relationship between two variables as points.
It is not a table that tabulates the combinations of predicted and actual class by count, so it is incorrect.
Confusion matrix
Correct. A confusion matrix organizes the combinations of actual class and predicted class (true positive, false positive, true negative, false negative) into a table. It reveals where errors are occurring and is the basis for calculating each metric.
Correlation matrix
A correlation matrix is a table that arranges the correlation coefficients between features.
It is a table of relationships between variables, not the breakdown of correct and incorrect predictions, so it is incorrect.
Remember the correct answer, the confusion matrix.
・It organizes the combinations of actual class and predicted class (true positive, false positive, true negative, false negative) into a table.
・It reveals where errors are occurring and is the basis for calculating metrics such as accuracy, precision, and recall.