Download presentation
Presentation is loading. Please wait.
Published byΚρέων Μέλιοι Modified over 5 years ago
1
Information Organization: Evaluation of Classification Performance
2
Classification: Evaluation Measures
Sensitivity = A / (A+C) Probability that classifier result is correct given that an item belongs to T Similar to Recall Specificity = D / (B+D) Probability that classifier result is correct given that an item belongs to F Positive Predictive Value = A / (A+B) Probability that classifier result is correct given that it is T Similar to Precision Negative Predictive Value = D / (C+D) Probability that classifier result is correct given that it is F Accuracy = (A+D) / (A+B+C+D) Probability that classifier result is correct F-measure Actual Predicted T F A B C D A = True Positive B = False Postive C = False Negative D = True Negative Search Engine
3
Classifier Evaluation: Examples
Senario 1 A ball sorter (classifier) with 99% sensitivity and 98% specificity puts 1000 green and 1000 white tennis balls into T-bin (w/ color) and F-bin (w/o color). A= 990 green balls in T-bin (true positive) C= 10 green balls in F-bin (false negative) D= 980 white balls in F-bin (true negative) B= 20 white balls in T-bin (false positive) Positive Predictive Value = 990 / (990+20) = 90% Negative Predictive Value = 980 / (10+980) = 98.99% Accuracy = ( ) / 2000 = 98.5% F-score = (2*990)/( ) = 94.29% Scenario 2: A ball sorter (classifier) with 99% sensitivity and 98% specificity puts 100 green and 1900 white tennis balls into T-bin (w/ color) and F-bin (w/o color). A= 99 green balls in T-bin (true positive) C= 1 green balls in F-bin (false negative) D= 1862 white balls in F-bin (true negative) B= 38 white balls in T-bin (false positive) Positive Predictive Value = 99 / (99+38) = 72.26% Negative Predictive Value = 1862 / (1+1862) = 99.95% Accuracy = ( ) / 2000 = 98.05% F-score = (2*99)/( ) = 83.54% Actual Predicted T F A B C D Sensitivity = A / (A+C) Specificity = D / (B+D) PPV = A / (A+B) NPV = D / (C+D) Accuracy = (A+D) / (A+B+C+D) F-score = 2A / (A+C+A+B) Search Engine
4
Classifier Evaluation: Multiple Classes
Actual Predicted T F 35 3 2 60 Actual Predicted T F 65 360 35 40 Actual Predicted T F 100 363 37 Class 1: Accuracy = 95/100 = 0.95 Class 2: Accuracy = 105/500 = 0.21 Micro-average Table: Accuracy = 200/600 = 0.33 Multiple classes Macro-Averaging compute performance for each class, then average over all classes gives equal weight to each class can be unduly influenced by small categories e.g., accuracy = ( ) / 2 = 0.58 Micro-Averaging one confusion matrix for all classes gives equal weights to each item (e.g. document) can be dominated by large categories e.g., accuracy = 200/600 = 0.33 Search Engine
5
n-Fold Cross Validation
A technique for estimating classifier performance To reduce variability for accurate prediction of classifier performance n-fold cross validation Construct n sets of training/test data, where training portion = (n-1)/n of the whole training data test portion = 1/n of the whole training data (e.g. 3-fold cross validation) Compute the classification rate on each set For each classifier, assess the consistency of each classification rate if not consistent, problem w/ data or classifier compute the average of n classification rate to compare w/ other classifiers Search Engine
6
Harmonic Mean Harmonic Mean F-measure A type of average
e.g., average of rates Inverse of the mean of the inverse Count of numbers divided by sum of their reciprocals F-measure Harmonic mean of Precision & Recall Example: roundtrip speed A to B: x miles 50 miles/hr from A to B : x/50 hrs 20 miles/hr from B to A: x/20 hrs avg. speed =
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.