detDataGet
DET (Detection Error Tradeoff) data generation
Contents
Syntax
- [ths, fpr, fnr, th, fp, fn, gPrm1, gPrm2, a, b]=detDataGet(data1, data2, method, prior, showPlot);
Description
[ths, fpr, fnr, th, fp, fn, gPrm1, gPrm2, a, b]=detDataGet(data1, data2, method, prior, showPlot) returns the parameters for DET plot.
- data1: likelihood vector for negative cases (with lower likelihood)
- data2: likelihood vector for positive cases (with higher likelihood)
- method: method for computing threshold that minimizes FP+FN
- method=1: FP = (count of FP cases)/(count of all data); FN is defined similarly.
- method=2: FP = (count of FP cases)/(total negative case count); FN is defined similarly.
- th: threhold (using Baysian rule where priors are multiplied)
- fp: false positive error rate (using Baysian rule where priors are multiplied)
- fn: false negative error rate (using Baysian rule where priors are multiplied)
- gPrm1: mu and sigma of class 1
- gPrm2: mu and sigma of class 2
- a, b: the fitting parameters of y=1/(1+exp(-a*(x-b))) for class-2 conditional probability (priors are multiplied)
References
- A. Martin, G. Doddington, T. Kamm, M. Ordowski, and M. Przybocki, "The DET curve in assessment of detection task performance,", in Proceedings of Eurospeech, Rhodes, Greece, 1997, pp. 1895�V1898.
Example
data1=randn(1,50)/2+1;
data2=randn(1,25)+2.5;
opt=detDataGet('defaultOpt');
[ths, fpr, fnr]=detDataGet(data1, data2, opt, 1);
