Why did i get double TP ( true positive ) than the NAB source on github

I run HTM NUPIC ( python 2.7) with API algorithm.

i run with data https://github.com/numenta/NAB/blob/master/data/realKnownCause/machine_temperature_system_failure.csv

and have the result ( i capture from my file)

I have install & setup NAB on python 3.6 ( on another laptop)

and optimize, score, normalize on the result file above i got the supprise result. I got 11 TP ( but the TP of HTM in the github just 5)

so the question is :smile:

  1. My models give better TP ( although the FP is higer too)
  2. Or My models have problem?
  3. Can we know exactly how many TP in evey data file ( because every data in NAB are labeled ? ) .

@rhyolight @Balladeer @marty1885 @sheiser1 and others please help.

Thank you all very much.

anybody help ? thank you very much.

This may be similar to this issue I have been working on:

As I said in the github issue, I am not sure if this is a problem or not, because those CSV files are intermediary files used to calculate the final scores in results/final_scores.json. If you find that the final scores are incorrect, please report how you ran NAB so I can try to replicate the problem. But I am not very concerned with discrepancies in these intermediary CSV files.

1 Like

thank you @rhyolight. i just wonder that Numenta created the corpus ( they labeled the 58 datasets) so they know exactly the amount of abomals and the time of every abnomal. Do you have any infomation about that ?