Add arbitrary decoding functionality to process module.
This should accept a processed recording file (h5 or nwb), a binary regressor file (npz, generate with visiomode-analysis) and a mask, either as a file (npy/npz), an ABA area or an all option for all ABA areas.
The module should support logistic regression and LDA decoders.
It should output a JSON file with the following keys:
decoder_type => type of decoder
decoded_variable => variable decoded
test_strategy => test strategy used, eg 80:20 split
train_strategy => training strategy used, eg random_shuffled
run_date => date decoder was run
accuracy => accuracy of decoder
balanced accuracy => balanced accuracy of decoder
f2_score => decoder F2 score
confusion matrix => decoder confusion matrix
d_prime => decoder d'
Add arbitrary decoding functionality to
processmodule.This should accept a processed recording file (h5 or nwb), a binary regressor file (npz, generate with
visiomode-analysis) and a mask, either as a file (npy/npz), an ABA area or analloption for all ABA areas.The module should support logistic regression and LDA decoders.
It should output a JSON file with the following keys:
decoder_type=> type of decoderdecoded_variable=> variable decodedtest_strategy=> test strategy used, eg 80:20 splittrain_strategy=> training strategy used, eg random_shuffledrun_date=> date decoder was runaccuracy=> accuracy of decoderbalanced accuracy=> balanced accuracy of decoderf2_score=> decoder F2 scoreconfusion matrix=> decoder confusion matrixd_prime=> decoder d'