Pengantar Deteksi Anomali di R
Alastair Rushworth
Data Scientist
Pilih nilai ambang tinggi
high_score <- quantile(sat$score, probs = 0.99)
high_score
99%
0.6228078
Binerkan skor
sat$binary_score <- as.numeric(score >= high_score)
Membandingkan label asli dan skor yang dibinerkan
table(sat$label, sat$binary_score)
0 1
0 5729 3
1 15 56
Anomali teridentifikasi benar $\div$ Total anomali
table(sat$label, sat$binary_score)
0 1
0 5729 3
1 15 56
recall <- 56 / (15 + 56)
recall
0.7887324
Anomali teridentifikasi benar $\mathbf{\div}$ Total yang ditandai anomali
table(sat$label, sat$binary_score)
0 1
0 5729 3
1 15 56
precision <- 56 / (56 + 3)
precision
0.9491525
Pengantar Deteksi Anomali di R