Figure A2
The grouped confusion matrix compares four models: “R F”, “X G B”, “K N N”, and “L R”, with values color-coded according to the legend (black for R F, green for X G B, blue for K N N, and red for L R). The matrix is divided into four sections based on actual and predicted classes. The top-left section, labeled “Real Non-fatal predicted as Non-fatal”, represents true negatives. The values are R F 21, X G B 21, K N N 19, and L R 20. The top-right section, labeled “Real Non-fatal predicted as Fatal”, represents false positives. The values are R F 9, X G B 9, K N N 11, and L R 10. The bottom-left section, labeled “Real Fatal predicted as Non-fatal”, represents false negatives. The values are R F 13, X G B 16, K N N 16, and L R 13. The bottom-right section, labeled “Real Fatal predicted as Fatal”, represents true positives. The values are R F 31, X G B 28, K N N 28, and L R 31. The background shading differentiates the four quadrants, with neutral tones for correct classifications and red-tinted areas for misclassifications.Integrated confusion matrix of the prediction models. Source: Authors’ own work