12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152 |
- function [bestEpsilon bestF1] = selectThreshold(yval, pval)
- %SELECTTHRESHOLD Find the best threshold (epsilon) to use for selecting
- %outliers
- % [bestEpsilon bestF1] = SELECTTHRESHOLD(yval, pval) finds the best
- % threshold to use for selecting outliers based on the results from a
- % validation set (pval) and the ground truth (yval).
- %
- bestEpsilon = 0;
- bestF1 = 0;
- stepsize = (max(pval) - min(pval)) / 1000;
- for epsilon = min(pval):stepsize:max(pval)
- cvP = pval < epsilon;
- tp = sum(cvP & yval);
- fp = sum(cvP & (yval == 0));
- fn = sum((cvP == 0) & yval);
- prec = tp / (tp + fp);
- rec = tp / (tp + fn);
- F1 = 2 * prec * rec / (prec + rec);
- % ====================== YOUR CODE HERE ======================
- % Instructions: Compute the F1 score of choosing epsilon as the
- % threshold and place the value in F1. The code at the
- % end of the loop will compare the F1 score for this
- % choice of epsilon and set it to be the best epsilon if
- % it is better than the current choice of epsilon.
- %
- % Note: You can use predictions = (pval < epsilon) to get a binary vector
- % of 0's and 1's of the outlier predictions
- % =============================================================
- if F1 > bestF1
- bestF1 = F1;
- bestEpsilon = epsilon;
- end
- end
- end
|