PyCM Report

Dataset Type :

Note 1 : Recommended statistics for this type of classification highlighted in aqua

Note 2 : The recommendation system assumes the input is the result of classification over the entire dataset, not just a subset. If the confusion matrix is based on test data classification, the recommendation may not be valid.

Confusion Matrix :

Actual Predict
0 1 2
0 13 0 0
1 0 15 1
2 0 0 9

Overall Statistics :

95% CI (0.92279,1.02458)
ACC Macro 0.98246
ARI 0.92263
AUNP 0.98276
AUNU 0.98384
Bangdiwala B 0.9519
Bennett S 0.96053
CBA 0.94583
CSI 0.94583
Chi-Squared 70.0625
Chi-Squared DF 4
Conditional Entropy 0.14202
Cramer V 0.96014
Cross Entropy 1.55021
F1 Macro 0.9717
F1 Micro 0.97368
FNR Macro 0.02083
FNR Micro 0.02632
FPR Macro 0.01149
FPR Micro 0.01316
Gwet AC1 0.9609
Hamming Loss 0.02632
Joint Entropy 1.68902
KL Divergence 0.0032
Kappa 0.95979
Kappa 95% CI (0.88202,1.03756)
Kappa No Prevalence 0.94737
Kappa Standard Error 0.03968
Kappa Unbiased 0.95977
Krippendorff Alpha 0.9603
Lambda A 0.95455
Lambda B 0.95652
Mutual Information 1.42359
NIR 0.42105
NPV Macro 0.98551
NPV Micro 0.98684
Overall ACC 0.97368
Overall CEN 0.06054
Overall J (2.8375,0.94583)
Overall MCC 0.96082
Overall MCEN 0.09387
Overall RACC 0.34557
Overall RACCU 0.34591
P-Value 0.0
PPV Macro 0.96667
PPV Micro 0.97368
Pearson C 0.8052
Phi-Squared 1.84375
RCI 0.92022
RR 12.66667
Reference Entropy 1.54701
Response Entropy 1.5656
SOA1(Landis & Koch) Almost Perfect
SOA2(Fleiss) Excellent
SOA3(Altman) Very Good
SOA4(Cicchetti) Excellent
SOA5(Cramer) Very Strong
SOA6(Matthews) Very Strong
SOA7(Lambda A) Very Strong
SOA8(Lambda B) Very Strong
SOA9(Krippendorff Alpha) High
SOA10(Pearson C) Strong
Scott PI 0.95977
Standard Error 0.02597
TNR Macro 0.98851
TNR Micro 0.98684
TPR Macro 0.97917
TPR Micro 0.97368
Zero-one Loss 1

Class Statistics :

Class 0 1 2 Description
ACC 1.0 0.97368 0.97368 Accuracy
AGF 1.0 0.95711 0.98556 Adjusted F-score
AGM 1.0 0.97989 0.97521 Adjusted geometric mean
AM 0 -1 1 Difference between automatic and manual classification
AUC 1.0 0.96875 0.98276 Area under the ROC curve
AUCI Excellent Excellent Excellent AUC value interpretation
AUPR 1.0 0.96875 0.95 Area under the PR curve
BB 1.0 0.9375 0.9 Braun-Blanquet similarity
BCD 0.0 0.01316 0.01316 Bray-Curtis dissimilarity
BM 1.0 0.9375 0.96552 Informedness or bookmaker informedness
CEN 0 0.07991 0.11179 Confusion entropy
DOR None None None Diagnostic odds ratio
DP None None None Discriminant power
DPI None None None Discriminant power interpretation
ERR 0.0 0.02632 0.02632 Error rate
F0.5 1.0 0.98684 0.91837 F0.5 score
F1 1.0 0.96774 0.94737 F1 score - harmonic mean of precision and sensitivity
F2 1.0 0.94937 0.97826 F2 score
FDR 0.0 0.0 0.1 False discovery rate
FN 0 1 0 False negative/miss/type 2 error
FNR 0.0 0.0625 0.0 Miss rate or false negative rate
FOR 0.0 0.04348 0.0 False omission rate
FP 0 0 1 False positive/type 1 error/false alarm
FPR 0.0 0.0 0.03448 Fall-out or false positive rate
G 1.0 0.96825 0.94868 G-measure geometric mean of precision and sensitivity
GI 1.0 0.9375 0.96552 Gini index
GM 1.0 0.96825 0.98261 G-mean geometric mean of specificity and sensitivity
HD 0 1 1 Hamming distance
IBA 1.0 0.87891 0.99881 Index of balanced accuracy
ICSI 1.0 0.9375 0.9 Individual classification success index
IS 1.54749 1.24793 1.926 Information score
J 1.0 0.9375 0.9 Jaccard index
LS 2.92308 2.375 3.8 Lift score
MCC 1.0 0.94696 0.93218 Matthews correlation coefficient
MCCI Very Strong Very Strong Very Strong Matthews correlation coefficient interpretation
MCEN 0 0.125 0.1661 Modified confusion entropy
MK 1.0 0.95652 0.9 Markedness
N 25 22 29 Condition negative
NLR 0.0 0.0625 0.0 Negative likelihood ratio
NLRI Good Good Good Negative likelihood ratio interpretation
NPV 1.0 0.95652 1.0 Negative predictive value
OC 1.0 1.0 1.0 Overlap coefficient
OOC 1.0 0.96825 0.94868 Otsuka-Ochiai coefficient
OP 1.0 0.94143 0.95614 Optimized precision
P 13 16 9 Condition positive or support
PLR None None 29.0 Positive likelihood ratio
PLRI None None Good Positive likelihood ratio interpretation
POP 38 38 38 Population
PPV 1.0 1.0 0.9 Precision or positive predictive value
PRE 0.34211 0.42105 0.23684 Prevalence
Q None None None Yule Q - coefficient of colligation
QI None None None Yule Q interpretation
RACC 0.11704 0.1662 0.06233 Random accuracy
RACCU 0.11704 0.16638 0.0625 Random accuracy unbiased
TN 25 22 28 True negative/correct rejection
TNR 1.0 1.0 0.96552 Specificity or true negative rate
TON 25 23 28 Test outcome negative
TOP 13 15 10 Test outcome positive
TP 13 15 9 True positive/hit
TPR 1.0 0.9375 1.0 Sensitivity, recall, hit rate, or true positive rate
Y 1.0 0.9375 0.96552 Youden index
dInd 0.0 0.0625 0.03448 Distance index
sInd 1.0 0.95581 0.97562 Similarity index

Generated By PyCM Version 4.2