In [1]:
%matplotlib inline
from matplotlib import pylab as pl
import cPickle as pickle
import pandas as pd
import numpy as np
import os
import random
from collections import defaultdict

In [2]:
import sys
sys.path.append('..')

Read precomputed features

uncommoent the relevant pipeline in ../seizure_detection.py and run

cd ..
./doall data

In [28]:
FEATURES = 'gen-8.5_medianwindow1-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9'

In [29]:
from common.data import CachedDataLoader
cached_data_loader = CachedDataLoader('../data-cache')

In [30]:
def read_data(target, data_type):
    return cached_data_loader.load('data_%s_%s_%s'%(data_type,target,FEATURES),None)

Predict


In [43]:
RF = True
if RF:
    max_depth=7
    from sklearn.ensemble import RandomForestClassifier

    clf = RandomForestClassifier(n_estimators=3000, min_samples_split=1, bootstrap=False,max_depth=max_depth,
                                 n_jobs=-1) #min_samples_leaf=4
    
    with_weights = False
    PWEIGHT = 0.
    LWEIGHT = 0.
    suffix = 'maxdepth%d.'%max_depth
else:
    with_weights = False
    PWEIGHT = 0.
    LWEIGHT = 0.
    suffix = 'dbn500.epochs100.'
#     sys.path.append('/Users/udi/Downloads/github/cudamat') #slower!
    from sklearn import preprocessing
    from nolearn.dbn import DBN
    from sklearn.pipeline import Pipeline

    min_max_scaler = preprocessing.MinMaxScaler() # scale features to be [0..1] which is DBN requirement

    dbn = DBN(
        [-1, 300, -1], # first layer has size X.shape[1], hidden layer(s), last layer will have number of classes in y (2))
        learn_rates=0.3,
        learn_rate_decays=0.9,
        epochs=100,
        verbose=0,
        )

    clf = Pipeline([('min_max_scaler', min_max_scaler), ('dbn', dbn)])

split examples into segments, each from the same event in each CV-split we will take all examples from the same segment to either train or validate


In [44]:
def getsegments(pdata):
    segments = []
    start = 0
    last_l = 0
    for i,l in enumerate(pdata.latencies):
        if l<last_l:
            segments.append(np.arange(start,i))
            start = i
        last_l = l
    segments.append(np.arange(start,i+1))
    return np.array(segments)

Compute AUC for each target separatly


In [45]:
%%time
import itertools
from sklearn.metrics import roc_auc_score

target2iter2ys = {}
for target in ['Dog_1', 'Dog_2', 'Dog_3', 'Dog_4', 'Dog_5', 'Patient_1', 'Patient_2']:
    # positive examples
    pdata = read_data(target, 'preictal')
    Np, NF = pdata.X.shape
    
    if not RF:
        clf.set_params(dbn__layer_sizes=[NF,300,2]) # we need to reset each time because NF is different
    
    psegments = getsegments(pdata)
    Nps = len(psegments)

    # negative examples
    ndata = read_data(target, 'interictal')
    Nn, NF = ndata.X.shape
    nsegments = getsegments(ndata)
    Nns = len(nsegments)
    
    npratio = float(Nn)/Np
    print target,1/(1+npratio),Np,Nn
    npsratio = float(Nns)/Nps
    print target,1/(1+npsratio),Nps,Nns
    Ntrainps = 1
    Ntrainns = int(Ntrainps*npsratio)

    iter2ys = defaultdict(list) # {niter: Ns *[(ytest,y_proba)]
    for s in psegments:
        Xtestp = pdata.X[s,:]
        weightstest = pdata.latencies[s] # latency for first segment is 1
        
        Ntrainp = len(s)
        Ntrainn = int(Ntrainp*npratio)
        
        good_iter = 0
        for niter in range(10):
#             n = np.array(random.sample(xrange(Nn),Ntrainn)) # segment based
            ns = np.array(random.sample(xrange(Nns),Ntrainns)) # sequence based
            n = np.array(list(itertools.chain(*nsegments[ns]))) # .ravel does not work - elements of nsegments are not of equal length
            Xtestn = ndata.X[n,:]

            Xtrainp = pdata.X[-s,:]
            Xtrainn = ndata.X[-n,:]

            Xtrain = np.concatenate((Xtrainp,Xtrainn))
            ytrain = np.concatenate((np.ones(Xtrainp.shape[0]),np.zeros(Xtrainn.shape[0])))
            perm = np.random.permutation(len(ytrain))
            ytrain = ytrain[perm]
            Xtrain = Xtrain[perm,:]

            Xtest = np.concatenate((Xtestp,Xtestn))
            ytest = np.concatenate((np.ones(Xtestp.shape[0]),np.zeros(Xtestn.shape[0])))

            if with_weights:
                weightsp = PWEIGHT*np.ones(Xtrainp.shape[0])
                weightsp += LWEIGHT * (pdata.latencies[-s]-1.) # latency for first segment is 1
                weightsn = np.ones(Xtrainn.shape[0]) 
                weights = np.concatenate((weightsp,weightsn))
                weights = weights[perm]

            if not RF:
                minibatch_size = 64
                if Xtrain.shape[0] < minibatch_size:
                    minibatch_size = Xtrain.shape[0]
                clf.set_params(dbn__minibatch_size=minibatch_size)

            if with_weights:
                clf.fit(Xtrain, ytrain, sample_weight=weights)
            else:
                clf.fit(Xtrain, ytrain)
            good_iter += 1

            y_proba = clf.predict_proba(Xtest)[:,1]
            iter2ys[good_iter].append((ytest, y_proba))
            
            auc = roc_auc_score(ytest, y_proba)
            print '%.3f'%auc,Ntrainp,np.mean(weightstest)
            if good_iter >= 2:
                break
    target2iter2ys[target] = iter2ys
    print


Dog_1 0.047619047619 184 3680
Dog_1 0.047619047619 4 80
0.740 46 3.5
0.868 46 3.5
0.541 46 3.5
0.415 46 3.5
0.524 46 3.5
0.630 46 3.5
0.539 46 3.5
0.492 46 3.5

Dog_2 0.0775903614458 322 3828
Dog_2 0.0769230769231 7 84
0.983 46 3.5
0.983 46 3.5
0.981 46 3.5
0.958 46 3.5
0.965 46 3.5
0.946 46 3.5
1.000 46 3.5
1.000 46 3.5
0.981 46 3.5
0.973 46 3.5
1.000 46 3.5
0.997 46 3.5
0.689 46 3.5
0.684 46 3.5

Dog_3 0.047619047619 552 11040
Dog_3 0.047619047619 12 240
0.946 46 3.5
0.910 46 3.5
0.827 46 3.5
0.817 46 3.5
0.803 46 3.5
0.952 46 3.5
0.388 46 3.5
0.716 46 3.5
0.552 46 3.5
0.547 46 3.5
1.000 46 3.5
1.000 46 3.5
1.000 46 3.5
1.000 46 3.5
0.976 46 3.5
0.975 46 3.5
0.477 46 3.5
0.417 46 3.5
0.835 46 3.5
0.663 46 3.5
0.879 46 3.5
0.844 46 3.5
0.315 46 3.5
0.476 46 3.5

Dog_4 0.106796116505 737 6164
Dog_4 0.112582781457 17 134
0.684 46 3.5
0.971 46 3.5
0.968 46 3.5
0.732 46 3.5
0.344 46 3.5
0.483 46 3.5
0.751 37 4.0
0.577 37 4.0
0.632 19 5.0
0.554 19 5.0
0.942 46 3.5
0.408 46 3.5
0.681 46 3.5
0.882 46 3.5
0.751 37 4.0
0.597 37 4.0
0.632 46 3.5
0.393 46 3.5
0.413 46 3.5
0.232 46 3.5
0.703 46 3.5
0.764 46 3.5
0.649 46 3.5
0.186 46 3.5
0.778 46 3.5
0.463 46 3.5
0.349 46 3.5
0.357 46 3.5
0.624 46 3.5
0.525 46 3.5
0.518 46 3.5
0.472 46 3.5
0.600 46 3.5
0.257 46 3.5

Dog_5 0.0625 230 3450
Dog_5 0.0625 5 75
0.867 46 3.5
0.800 46 3.5
0.899 46 3.5
0.757 46 3.5
1.000 46 3.5
1.000 46 3.5
0.996 46 3.5
0.974 46 3.5
1.000 46 3.5
0.936 46 3.5

Patient_1 0.267441860465 138 378
Patient_1 0.25 3 9
0.734 46 3.5
0.732 46 3.5
1.000 46 3.5
1.000 46 3.5
0.264 46 3.5
0.736 46 3.5

Patient_2 0.3 138 322
Patient_2 0.3 3 7
1.000 46 3.5
0.756 46 3.5
1.000 46 3.5
1.000 46 3.5
0.961 46 3.5
0.937 46 3.5

CPU times: user 20min 43s, sys: 29 s, total: 21min 11s
Wall time: 5min 12s

In [46]:
fname = '../data-cache/140926-CV.%s%s.pkl'%(suffix, FEATURES)
with open(fname,'wb') as fp:
    pickle.dump(target2iter2ys,fp,-1)

In [47]:
fname


Out[47]:
'../data-cache/140926-CV.maxdepth7.gen-8.5_medianwindow1-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9.pkl'

Generate a single AUC score


In [48]:
from sklearn.metrics import roc_auc_score
def p(a,b):
    return '%d E%d'%(1000*a,1000*b)

for f in [
'maxdepth7.gen-8.5_medianwindow1-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9',
'maxdepth10.gen-8.5_medianwindow1-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9',
'maxdepth15.gen-8.5_medianwindow1-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9',
'maxdepth5.gen-8_medianwindow1-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9',
'maxdepth10.gen-8_medianwindow1-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9',
'maxdepth20.gen-8_medianwindow1-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9',
'maxdepth5.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9-pca',
'maxdepth10.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9-pca',
'maxdepth15.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9-pca',
'maxdepth20.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9-pca',
'dbn500.epochs100.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9',
'dbn300.epochs100.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9',
'dbn100.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9',
'dbn300.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9',
            'maxdepthNone.gen-8_medianwindow-bands2-usf-w10-hammingP2-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9',
            'maxdepth25.gen-8_medianwindow-bands2-usf-w10-hammingP2-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9',
            'maxdepth20.gen-8_medianwindow-bands2-usf-w10-hammingP2-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9',
            'maxdepth15.gen-8_medianwindow-bands2-usf-w10-hammingP2-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9',
            'max_depth5gen-8_medianwindow-bands2-usf-w10-hammingP2-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9',
            'max_depth10gen-8_medianwindow-bandsI2-usf-w60-b0.2-b4-b8-b12-b30-b50-b75-b100-b117-0.1-0.5-0.9',
            'max_depth5gen-8_medianwindow-bandsI2-usf-w60-b0.2-b4-b8-b12-b30-b50-b75-b100-b117-0.1-0.5-0.9',
            'max_depth5.min_samples_leaf4.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9',
            'maxdepth10.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9',
            'max_depth5.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9',
        ]:
    all_ytest = all_y_proba =None
    all_aucs = []
    with open('../data-cache/140926-CV.%s.pkl'%f,'rb') as fp:
        target2iter2ys = pickle.load(fp)
    for target, iter2ys in target2iter2ys.iteritems():
        target_ytest = target_y_proba =None
        target_aucs = []
        print target,
        for ys in iter2ys.itervalues():
            ytest = y_proba =None
            aucs = []
            for y in ys:
                yt, yp = y
                ytest = yt if ytest is None else np.concatenate((ytest,yt))
                y_proba = yp if y_proba is None else np.concatenate((y_proba,yp))
                aucs.append(roc_auc_score(yt, yp))
            print p(roc_auc_score(ytest, y_proba), np.mean(aucs)),
            target_aucs += aucs
            target_ytest = ytest if target_ytest is None else np.concatenate((target_ytest,ytest))
            target_y_proba = y_proba if target_y_proba is None else np.concatenate((target_y_proba,y_proba))
        print target,p(roc_auc_score(target_ytest, target_y_proba),np.mean(target_aucs))
        all_aucs += target_aucs        
        all_ytest = target_ytest if all_ytest is None else np.concatenate((all_ytest,target_ytest))
        all_y_proba = target_y_proba if all_y_proba is None else np.concatenate((all_y_proba,target_y_proba))
#         if target == 'Dog_3':
#             pl.hist(target_aucs,alpha=0.5)
    print f,p(roc_auc_score(all_ytest, all_y_proba),np.mean(all_aucs))
    print


Dog_2 888 E942 881 E934 Dog_2 884 E938
Dog_3 759 E749 764 E776 Dog_3 761 E763
Dog_1 593 E585 602 E601 Dog_1 597 E593
Dog_4 616 E648 543 E520 Dog_4 579 E584
Dog_5 951 E952 894 E893 Dog_5 922 E922
Patient_2 910 E987 917 E897 Patient_2 919 E942
Patient_1 777 E666 770 E822 Patient_1 773 E744
maxdepth7.gen-8.5_medianwindow1-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9 764 E739

Dog_2 897 E919 886 E892 Dog_2 891 E906
Dog_3 804 E778 717 E700 Dog_3 762 E739
Dog_1 649 E650 628 E620 Dog_1 638 E635
Dog_4 633 E671 632 E639 Dog_4 632 E655
Dog_5 945 E963 941 E963 Dog_5 943 E963
Patient_2 860 E965 970 E966 Patient_2 918 E966
Patient_1 767 E772 762 E846 Patient_1 763 E809
maxdepth10.gen-8.5_medianwindow1-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9 783 E765

Dog_2 898 E937 883 E898 Dog_2 889 E917
Dog_3 742 E725 737 E723 Dog_3 740 E724
Dog_1 646 E637 660 E645 Dog_1 653 E641
Dog_4 672 E703 624 E628 Dog_4 647 E666
Dog_5 948 E960 890 E904 Dog_5 920 E932
Patient_2 752 E837 778 E722 Patient_2 755 E779
Patient_1 767 E704 779 E782 Patient_1 773 E743
maxdepth15.gen-8.5_medianwindow1-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9 777 E749

Dog_2 908 E944 895 E922 Dog_2 902 E933
Dog_3 750 E758 703 E711 Dog_3 726 E735
Dog_1 599 E595 572 E555 Dog_1 587 E575
Dog_4 609 E640 591 E625 Dog_4 599 E633
Dog_5 917 E905 916 E914 Dog_5 917 E909
Patient_2 797 E698 995 E989 Patient_2 922 E844
Patient_1 865 E819 763 E660 Patient_1 804 E739
maxdepth5.gen-8_medianwindow1-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9 762 E739

Dog_2 887 E932 847 E900 Dog_2 866 E916
Dog_3 714 E716 801 E790 Dog_3 757 E753
Dog_1 669 E659 671 E690 Dog_1 670 E674
Dog_4 632 E647 637 E635 Dog_4 634 E641
Dog_5 898 E914 923 E928 Dog_5 911 E921
Patient_2 740 E759 689 E710 Patient_2 717 E735
Patient_1 882 E812 832 E751 Patient_1 853 E782
maxdepth10.gen-8_medianwindow1-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9 777 E749

Dog_2 889 E919 908 E954 875 E914 Dog_2 890 E929
Dog_3 794 E818 766 E787 771 E777 Dog_3 777 E794
Dog_1 605 E605 538 E524 603 E605 Dog_1 581 E578
Dog_4 595 E610 565 E569 556 E583 Dog_4 571 E587
Dog_5 937 E926 908 E901 947 E951 Dog_5 930 E926
Patient_2 991 E992 977 E949 858 E973 Patient_2 936 E972
Patient_1 819 E807 886 E866 695 E636 Patient_1 797 E770
maxdepth20.gen-8_medianwindow1-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9 767 E749

Dog_2 863 E918 881 E926 Dog_2 872 E922
Dog_3 838 E830 813 E811 Dog_3 826 E820
Dog_1 634 E644 513 E496 Dog_1 574 E570
Dog_4 573 E572 535 E526 Dog_4 553 E549
Dog_5 900 E915 873 E879 Dog_5 886 E897
Patient_2 767 E855 789 E782 Patient_2 774 E818
Patient_1 711 E740 690 E760 Patient_1 704 E750
maxdepth5.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9-pca 746 E727

Dog_2 897 E931 871 E928 Dog_2 884 E930
Dog_3 836 E833 841 E840 Dog_3 839 E837
Dog_1 541 E511 559 E555 Dog_1 553 E533
Dog_4 518 E517 551 E580 Dog_4 535 E548
Dog_5 924 E953 936 E964 Dog_5 930 E959
Patient_2 880 E935 818 E946 Patient_2 852 E941
Patient_1 624 E542 676 E719 Patient_1 659 E630
maxdepth10.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9-pca 763 E735

Dog_2 904 E960 880 E908 Dog_2 891 E934
Dog_3 867 E858 864 E854 Dog_3 865 E856
Dog_1 554 E533 535 E528 Dog_1 545 E531
Dog_4 532 E530 542 E553 Dog_4 537 E541
Dog_5 918 E935 875 E862 Dog_5 899 E899
Patient_2 716 E773 725 E802 Patient_2 721 E788
Patient_1 623 E626 532 E445 Patient_1 580 E536
maxdepth15.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9-pca 765 E718

Dog_2 860 E908 904 E944 Dog_2 882 E926
Dog_3 794 E787 858 E853 Dog_3 828 E820
Dog_1 543 E540 542 E553 Dog_1 542 E547
Dog_4 584 E595 580 E575 Dog_4 581 E585
Dog_5 934 E936 887 E919 Dog_5 912 E928
Patient_2 662 E683 869 E859 Patient_2 766 E771
Patient_1 605 E687 540 E545 Patient_1 571 E616
maxdepth20.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9-pca 759 E730

Dog_2 744 E749 830 E833 Dog_2 785 E791
Dog_3 621 E642 630 E667 Dog_3 624 E655
Dog_1 440 E440 382 E378 Dog_1 412 E409
Dog_4 573 E604 602 E651 Dog_4 588 E628
Dog_5 829 E875 765 E824 Dog_5 800 E850
Patient_2 439 E551 560 E586 Patient_2 492 E569
Patient_1 739 E604 911 E864 Patient_1 824 E734
dbn500.epochs100.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9 619 E664

Dog_2 751 E793 768 E806 Dog_2 755 E800
Dog_3 614 E618 633 E677 Dog_3 625 E647
Dog_1 402 E415 444 E481 Dog_1 426 E448
Dog_4 575 E627 551 E534 Dog_4 563 E580
Dog_5 781 E823 772 E814 Dog_5 778 E818
Patient_2 493 E519 422 E549 Patient_2 460 E534
Patient_1 746 E665 858 E820 Patient_1 798 E742
dbn300.epochs100.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9 588 E646

Dog_2 557 E661 635 E710 Dog_2 584 E685
Dog_3 596 E671 572 E668 Dog_3 585 E669
Dog_1 465 E408 467 E429 Dog_1 465 E418
Dog_4 527 E525 519 E524 Dog_4 525 E524
Dog_5 675 E760 629 E667 Dog_5 649 E714
Patient_2 692 E745 380 E142 Patient_2 525 E443
Patient_1 825 E863 793 E825 Patient_1 805 E844
dbn100.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9 603 E605

Dog_2 687 E746 699 E740 Dog_2 679 E743
Dog_3 667 E662 543 E661 Dog_3 599 E661
Dog_1 448 E454 377 E299 Dog_1 425 E377
Dog_4 572 E587 576 E605 Dog_4 571 E596
Dog_5 520 E544 566 E706 Dog_5 541 E625
Patient_2 478 E383 677 E634 Patient_2 570 E508
Patient_1 788 E800 726 E680 Patient_1 757 E740
dbn300.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9 636 E620

Dog_2 879 E932 893 E953 894 E941 Dog_2 889 E942
Dog_3 759 E788 766 E776 774 E772 Dog_3 765 E779
Dog_1 627 E656 666 E654 672 E679 Dog_1 653 E663
Dog_4 579 E604 542 E575 594 E599 Dog_4 571 E593
Dog_5 896 E930 907 E925 910 E894 Dog_5 905 E917
Patient_2 592 E580 836 E753 736 E878 Patient_2 723 E737
Patient_1 853 E833 874 E855 847 E780 Patient_1 844 E823
maxdepthNone.gen-8_medianwindow-bands2-usf-w10-hammingP2-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9 762 E744

Dog_2 911 E955 909 E938 881 E917 Dog_2 901 E937
Dog_3 748 E755 734 E731 746 E741 Dog_3 743 E742
Dog_1 662 E652 533 E540 661 E653 Dog_1 619 E615
Dog_4 596 E592 545 E541 587 E589 Dog_4 575 E574
Dog_5 960 E953 928 E931 937 E941 Dog_5 942 E942
Patient_2 980 E972 757 E842 846 E990 Patient_2 865 E935
Patient_1 811 E759 890 E928 760 E883 Patient_1 820 E856
maxdepth25.gen-8_medianwindow-bands2-usf-w10-hammingP2-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9 768 E740

Dog_2 906 E944 877 E931 885 E932 Dog_2 888 E936
Dog_3 774 E783 779 E769 768 E767 Dog_3 774 E773
Dog_1 680 E690 611 E605 672 E694 Dog_1 654 E663
Dog_4 579 E606 620 E653 634 E633 Dog_4 610 E631
Dog_5 937 E965 907 E904 944 E954 Dog_5 928 E941
Patient_2 893 E972 764 E853 905 E981 Patient_2 837 E935
Patient_1 888 E897 839 E864 905 E872 Patient_1 873 E877
maxdepth20.gen-8_medianwindow-bands2-usf-w10-hammingP2-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9 782 E771

Dog_2 884 E927 898 E911 872 E911 Dog_2 885 E916
Dog_3 730 E738 718 E720 736 E746 Dog_3 728 E735
Dog_1 588 E607 609 E600 650 E657 Dog_1 618 E621
Dog_4 525 E532 529 E524 586 E612 Dog_4 546 E556
Dog_5 895 E883 893 E904 914 E921 Dog_5 901 E903
Patient_2 847 E950 905 E831 894 E906 Patient_2 880 E896
Patient_1 803 E807 882 E816 911 E881 Patient_1 862 E835
maxdepth15.gen-8_medianwindow-bands2-usf-w10-hammingP2-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9 753 E723

Dog_2 887 E921 919 E955 895 E922 Dog_2 899 E933
Dog_3 733 E758 738 E760 738 E767 Dog_3 737 E762
Dog_1 612 E615 598 E622 618 E638 Dog_1 611 E625
Dog_4 592 E600 599 E619 537 E551 Dog_4 577 E590
Dog_5 939 E943 940 E946 913 E923 Dog_5 931 E937
Patient_2 894 E934 840 E847 914 E980 Patient_2 884 E920
Patient_1 705 E644 778 E765 757 E709 Patient_1 747 E706
max_depth5gen-8_medianwindow-bands2-usf-w10-hammingP2-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9 763 E740

Dog_2 885 E910 894 E927 888 E915 Dog_2 888 E917
Dog_3 744 E751 767 E773 722 E723 Dog_3 744 E749
Dog_1 621 E585 706 E711 672 E663 Dog_1 667 E653
Dog_4 516 E502 581 E590 533 E563 Dog_4 545 E552
Dog_5 929 E929 912 E916 946 E961 Dog_5 927 E935
Patient_2 754 E708 916 E883 806 E788 Patient_2 821 E793
Patient_1 709 E686 651 E604 850 E892 Patient_1 734 E727
max_depth10gen-8_medianwindow-bandsI2-usf-w60-b0.2-b4-b8-b12-b30-b50-b75-b100-b117-0.1-0.5-0.9 755 E718

Dog_2 872 E912 854 E889 864 E895 Dog_2 864 E898
Dog_3 790 E798 768 E777 742 E765 Dog_3 766 E780
Dog_1 676 E682 640 E651 589 E606 Dog_1 635 E646
Dog_4 591 E611 636 E684 531 E533 Dog_4 585 E609
Dog_5 961 E967 927 E918 946 E934 Dog_5 945 E940
Patient_2 968 E953 808 E972 784 E684 Patient_2 850 E870
Patient_1 637 E667 856 E857 907 E860 Patient_1 779 E795
max_depth5gen-8_medianwindow-bandsI2-usf-w60-b0.2-b4-b8-b12-b30-b50-b75-b100-b117-0.1-0.5-0.9 765 E751

Dog_2 861 E895 861 E909 879 E916 Dog_2 866 E907
Dog_3 716 E750 753 E756 738 E714 Dog_3 734 E740
Dog_1 617 E628 552 E562 563 E594 Dog_1 575 E595
Dog_4 617 E643 591 E605 540 E567 Dog_4 581 E605
Dog_5 905 E893 902 E923 957 E958 Dog_5 920 E925
Patient_2 868 E958 792 E727 676 E668 Patient_2 797 E784
Patient_1 722 E735 872 E862 706 E694 Patient_1 766 E764
max_depth5.min_samples_leaf4.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9 750 E729

Dog_2 861 E895 861 E909 879 E916 Dog_2 866 E907
Dog_3 716 E750 753 E756 738 E714 Dog_3 734 E740
Dog_1 617 E628 552 E562 563 E594 Dog_1 575 E595
Dog_4 617 E643 591 E605 540 E567 Dog_4 581 E605
Dog_5 905 E893 902 E923 957 E958 Dog_5 920 E925
Patient_2 868 E958 792 E727 676 E668 Patient_2 797 E784
Patient_1 722 E735 872 E862 706 E694 Patient_1 766 E764
maxdepth10.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9 750 E729

Dog_2 889 E919 908 E954 875 E914 Dog_2 890 E929
Dog_3 794 E818 766 E787 771 E777 Dog_3 777 E794
Dog_1 605 E605 538 E524 603 E605 Dog_1 581 E578
Dog_4 595 E610 565 E569 556 E583 Dog_4 571 E587
Dog_5 937 E926 908 E901 947 E951 Dog_5 930 E926
Patient_2 991 E992 977 E949 858 E973 Patient_2 936 E972
Patient_1 819 E807 886 E866 695 E636 Patient_1 797 E770
max_depth5.gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9 767 E749


In [24]: