 |
 |
Dear Jianhang Zhou, the classifier used in FIX internally calculates a probability for a component of being signal or noise, but then needs to decide on the final answer (remove/don’t remove) and this is done according to the threshold. The numbers provided as possible thresholds are not probabilities, but related to that. The lower the threshold, the more likely it will be for a component to be classified as signal (hence increased TPR, but also means some will not be true signals, so also lower TNR). The higher the threshold, the more likely a component will be classified as noise (higher TNR, but possibly removing some signal so lower TPR) Hope it’s clearer now. Best wishes, Ludovica > On 30 Oct 2018, at 6:04 pm, Jianhang Zhou <[log in to unmask]> wrote: > > I found a explanation (https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=fsl;860e8150.1607) for fix4melview interpretation, given by Dr. Ludovica Griffanti. > > Citation: > "for each component, after the number, you have the classification provided by FIX (signal, Unknown, Unclassified noise). > The true/false after that refers to what FIX would do when removing the bad components: False = not to be removed, True = To be removed." > > I didn't find answers for thresholds. What does the threshold indicate? For example, what would be different if thr 1 or thr 50 is applied, besides TPR and TNR? > > Thank you! > > Jianhang Zhou > > ######################################################################## > > To unsubscribe from the FSL list, click the following link: > https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1 ######################################################################## To unsubscribe from the FSL list, click the following link: https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1