Table 2.
(a) | |||||
---|---|---|---|---|---|
Target | Criterion Function Results | ||||
Predictor A | 0 | 1 | q = 0.5 | q = 1 | q = 2.5 |
0 | 18 | 23 | 0.108 | 0.090 | 0.056 |
1 | 278 | 0 | 0 | 0 | 0 |
mean conditional entropy | 0.108* | 0.090* | 0.056 | ||
(b) | |||||
Target | Criterion Function Results | ||||
Predictor B | 0 | 1 | q = 0.5 | q = 1 | q = 2.5 |
0 | 1 | 16 | 0.024 | 0.013 | 0.005 |
1 | 295 | 7 | 0.265 | 0.104 | 0.036 |
mean conditional entropy | 0.289 | 0.117 | 0.041* |
Example of change on the inferred predictor by using different values for q entropic parameter: (a) distributions that lead to wrong predictor and (b) distributions that lead to correct predictor.