Prepearing data



In [1]:

    
import numpy as np
import pandas as pd
import matplotlib.pyplot as pl
import shap

import colors
# load JS visualization code to notebook
shap.initjs()



In [2]:

    
data_features = pd.read_hdf('/Users/OlgaKo/Desktop/from_lab14/data_features_JetHT.hdf5', "data").reset_index(drop=True)
labels = 1-pd.read_hdf('/Users/OlgaKo/Desktop/from_lab14/labels_JetHT.hdf5', 'labels').reset_index(drop=True)
sublabels = pd.read_hdf('/Users/OlgaKo/Desktop/from_lab14/data_not_f_JetHT.hdf5', "data").reset_index(drop=True)



In [3]:

    
data_features.shape, sublabels.shape









    Out[3]:





((162990, 2470), (162990, 20))

some runs with given cause of anomaly



In [61]:

    
dict_causes = {
 273150:" [[61, 64], [66, 75]],     Ecal excluded",
 273296:" [[1, 94]],    Pixe excluded",
 273426:" [[1, 59], [64, 64], [66, 69]],inefficeincy in EE",
 273445:" [[5, 6]],  tracker off  seen by many systems",
 274142:" [[99, 100]],  pixel and tracker off",
 274157:" [[103, 534]],  EB minus has region with low efficiency",
 274282:" [[86, 88]],  tracker is off --> particle flow object are affected",
 275326:" [[1, 169]],  sistrip excl",
 275757:" [[104, 122]],  low DCS: HBHE. nothing on PF plot",
 275758:" [[1, 4]],   part of hcal not available --> JetMET reconstruction suffers",
 275764:" [[1, 31]],   pixel off",
 275766:" [[1, 23], [25, 60]],  lower tracker efficiency",
 275768:" [[1, 79]],  lower tracker efficiency",
 275769:" [[1, 22]],  pixel not in Data Acquisition",
 275781:" [[1, 29]],  EB+16, +17, +18 are excluded due to a problem with VME crateS206h",
 275783:" [[1, 1634]],  strip EXCL",
 275838:" [[1, 51]],  strip EXCL",
 275922:" [[4, 6]],  low stat",
 276064:" [[1, 22]],  hot trigger tower seen in EE-07 (high occupancy and Et TP) which cause  ~100% deadtime",
 276071:" [[1, 22]],   strip EXCL",
 276095:" [[1, 5]],   low stat",
 276237:" [[1, 5]],   ECAL, HCAL, PIXEL excluded",
 276453:" [[1, 8], [10, 125]],  EB-17 (FED 626): was excluded (because of cooling failure)",
 276455:" [[1, 401]],  strip EXCL",
 276456:" [[1, 182]],  strip EXCL",
 276457:" [[1, 19]],  sistrip not in DAQ",
 277217:" [[12, 14]],  Short collision run with strip in error in DCS and pixel in error in DAQ. For physically meaningful lumisections 33-47, the total rate is zero. L1T flags marked as bad due to this",
 277933:" [[1, 42]],  ECAL EE+09 FED 648 removed from all LS in the run because it was causing 100% dead time. EE+09 was not masked, so all LS in this run are bad",
 278309:" [[1, 10]],  EE-04 FED 607 TT disabled in LS [1-10] according to express dataset (LS# 10)",
 278821:" [[1, 33], [36, 37]],  FED652 in error, EE+04 is off",
 279028:" [[1, 17]],  strip EXCL",
 279995:" [[1, 8]],  hcal water colling issues",
 280002:" [[1, 111]],  ecal excluded",
 280006:" [[1, 68]],  EB-11 token ring missing",
 280007:" [[1, 36]],  Low voltage channel broken in EB",
 280239:" [[1, 9]],  strip EXCL",
 280241:" [[1, 11]],  strip EXCL",
 281663:" [[61, 172]],  Timing shifted by 5ns due to TCDS* problem ",
 281674:" [[1, 45]],  Timing shifted by 5ns due to TCDS* problem",
 281680:" [[1, 31]],  Timing shifted by 5ns due to TCDS* problem",
 281974:" [[79, 85]],  pixel High voltage OFF",
 282408:" [[80, 191]],  strip EXCL",
 282707:" [[79, 81]],  pixel High voltage OFF",
 282796:" [[80, 82]],  pixel High Voltage off",
 282921:" [[1, 152]],       strip EXCL"

    
}



In [5]:

    
indx_train = np.arange(data_features.shape[0]-int(data_features.shape[0]/5), dtype='int32')
indx_test = np.arange(data_features.shape[0]-int(data_features.shape[0]/5),data_features.shape[0], dtype='int32')



In [6]:

    
indx_known = sublabels[sublabels["runId"].isin(dict_causes.keys())].index.tolist()
ids_known = sublabels[['runId', 'lumiId']].iloc[indx_known]



In [7]:

    
indx_train_minus = [s for s in indx_train if (s not in indx_known)]



In [8]:

    
num_good = np.sum(labels)
num_bad = len(labels)-np.sum(labels)

weights = 0.5 / np.where(labels == 0.0, num_good, num_bad)
weights *= len(labels)



In [9]:

    
y_train = np.array(labels.iloc[indx_train_minus], 'float32')
y_known = np.array(labels.iloc[indx_known], 'float32')
y_test = np.array(labels.iloc[indx_test], 'float32')

X_train = np.array(data_features.iloc[indx_train_minus], 'float32')
X_known = np.array(data_features.iloc[indx_known], 'float32')
X_test = np.array(data_features.iloc[indx_test], 'float32')

weights_train = weights[indx_train_minus]
weights_test = weights[indx_test]



In [10]:

    
feature_names = data_features.columns



In [11]:

    
Muon_features = [s for s in feature_names if (s[:3] == 'qMu')]# and (s[3:7] != 'Cosm')]
Pho_features = [s for s in feature_names if s[:4] == 'qPho']
Cal_features = [s for s in feature_names if s[:4] == 'qCal']
PF_features = [s for s in feature_names if s[:3] == 'qPF']

channels_features = dict()

channels_features['muons'] = Muon_features
channels_features['photons'] = Pho_features
channels_features['PF'] = PF_features
channels_features['calo'] = Cal_features



In [12]:

    
num_fch = dict()
for g, fs in channels_features.items():
    num_fch[g]=len(fs)



In [13]:

    
num_fch









    Out[13]:





{'PF': 878, 'calo': 280, 'muons': 439, 'photons': 224}



In [14]:

    
vmins = []
vmaxs = []

for f in data_features.columns:
    values = data_features[f].values
    # trim the color range, but prevent the color range from collapsing
    vmin = np.nanpercentile(values, 5)
    vmax = np.nanpercentile(values, 95)
    if vmin == vmax:
        vmin = np.nanpercentile(values, 1)
        vmax = np.nanpercentile(values, 99)
        if vmin == vmax:
            vmin = np.min(values)
            vmax = np.max(values)
    vmins.append(vmin)
    vmaxs.append(vmax)



In [51]:

    
vmins = np.array(vmins)
vmaxs = np.array(vmaxs)



In [15]:

    
X_ref=X_test[np.random.choice(np.where(y_test==1)[0], size=100, replace=False)]

train any classifier



In [16]:

    
import xgboost as xgb
xgb_clf  = xgb.XGBClassifier(n_estimators=80, max_depth=10, learning_rate=0.1, silent=False)
xgb_clf.fit(X_train, y_train)









    Out[16]:





XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,
       colsample_bytree=1, gamma=0, learning_rate=0.1, max_delta_step=0,
       max_depth=10, min_child_weight=1, missing=None, n_estimators=80,
       n_jobs=1, nthread=None, objective='binary:logistic', random_state=0,
       reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,
       silent=False, subsample=1)



In [18]:

    
model = xgb_clf

SHAP to explain predictions



In [22]:

    
shap_values = shap.TreeExplainer(model).shap_values(X_known)
shap_ref = shap.TreeExplainer(model).shap_values(X_ref)



In [69]:

    
def summary_plot_custom(shap_values, features=None, feature_names=None, max_display=20,
                 vmins=None, vmaxs=None, return_oder=True,
                 axis_color="#333333", title=None, alpha=1, show=True, sort=True,
                 color_bar=True, auto_size_plot=True):

    num_features = shap_values.shape[1]


    if sort:
        # order features by the sum of their effect magnitudes
        feature_order = np.argsort(np.sum(np.abs(shap_values), axis=0))
        feature_order = feature_order[-min(max_display, len(feature_order)):]
    else:
        feature_order = np.arange(min(max_display, num_features))

        
    row_height = 0.4
    if auto_size_plot:
        pl.gcf().set_size_inches(8, len(feature_order) * row_height + 1.5)
    pl.axvline(x=0, color="#999999", zorder=-1)


    for pos, i in enumerate(feature_order):
        pl.axhline(y=pos, color="#cccccc", lw=0.5, dashes=(1, 5), zorder=-1)
        shaps = shap_values[:, i]
        values = features[:, i]


        N = len(shaps)
        nbins = 100
        quant = np.round(nbins * (shaps - np.min(shaps)) / (np.max(shaps) - np.min(shaps) + 1e-8))
        inds = np.argsort(quant + np.random.randn(N) * 1e-6)
        layer = 0
        last_bin = -1
        ys = np.zeros(N)
        for ind in inds:
            if quant[ind] != last_bin:
                layer = 0
            ys[ind] = np.ceil(layer / 2) * ((layer % 2) * 2 - 1)
            layer += 1
            last_bin = quant[ind]
        ys *= 0.9 * (row_height / np.max(ys + 1))

        vmin=vmins[i]
        vmax=vmaxs[i]

        nan_mask = np.isnan(values)
        
        pl.scatter(shaps[nan_mask], pos + ys[nan_mask], color="#777777", vmin=vmin,
                   vmax=vmax, s=16, alpha=alpha, linewidth=0,
                   zorder=3, rasterized=len(shaps) > 500)
        
        
        pl.scatter(shaps[np.invert(nan_mask)], pos + ys[np.invert(nan_mask)],
                   cmap=colors.red_blue, vmin=vmin, vmax=vmax, s=16,
                   c=values[np.invert(nan_mask)], alpha=alpha, linewidth=0,
                   zorder=3, rasterized=len(shaps) > 500)



    import matplotlib.cm as cm
    m = cm.ScalarMappable(cmap=pl.get_cmap(red_blue))
    m.set_array([0, 1])
    cb = pl.colorbar(m, ticks=[0, 1], aspect=1000)
    cb.set_ticklabels(['feature value low', 'feature value high'])
    cb.set_label('feature value', size=12, labelpad=0)
    cb.ax.tick_params(labelsize=11, length=0)
    cb.set_alpha(1)
    cb.outline.set_visible(False)
    bbox = cb.ax.get_window_extent().transformed(pl.gcf().dpi_scale_trans.inverted())
    cb.ax.set_aspect((bbox.height - 0.9) * 20)

    pl.gca().xaxis.set_ticks_position('bottom')
    pl.gca().yaxis.set_ticks_position('none')
    pl.gca().spines['right'].set_visible(False)
    pl.gca().spines['top'].set_visible(False)
    pl.gca().spines['left'].set_visible(False)
    pl.gca().tick_params(color=axis_color, labelcolor=axis_color)
    pl.yticks(range(len(feature_order)), [feature_names[i] for i in feature_order], fontsize=13)
    pl.gca().tick_params('x', labelsize=11)
    pl.ylim(-1, len(feature_order))
    pl.xlabel('shap value', fontsize=13)
    if show:
        pl.show()
    
    if return_oder:
        return feature_order



In [70]:

    
for run in dict_causes.keys():
    if run in sublabels['runId'].values:
        print('')
        print (run, dict_causes[run])
        print('')
            
        inds = np.where(np.array(ids_known)==run)[0]

        print ("sum of features' influences from all lumis, 20 the most important")
        print('')
        sum_shap = np.sum(shap_values[inds,:-1], axis=0)
        print (feature_names[np.argsort(sum_shap, axis=0)][:20])
        print('')
        
        print("the most common features with big shap values")
        print('')
        top_sh_values = np.argsort(shap_values[inds,:-1], axis=1)[:,:50].T.reshape(-1)
        unique, first_index, counts = np.unique(
            top_sh_values,
        return_counts=True, return_index=True)
        print(feature_names[top_sh_values[np.sort(first_index[counts == len(inds)])]])
        print('')
        
        print ('summary plot')
        order=summary_plot_custom(shap_values[inds,:], features=X_known[inds, :], feature_names=feature_names,
                               vmins=vmins, vmaxs=vmaxs, return_oder=True, sort=True)
        
        print ('reference plot for good lumis')
        summary_plot_custom(shap_ref[:,order], features=X_ref[:,order], feature_names=feature_names[order],
                               vmins=vmins[order], vmaxs=vmaxs[order], return_oder=False, sort=False)
        
        channels_shap = pd.DataFrame()
        for i in shap_values[inds,:-1]:
            pred_for_channnel = {}
            for channel in channels_features:
                current_sum = 0
                for feature in channels_features[channel]:
                    current_sum += i[list(feature_names).index(feature)]
                pred_for_channnel[channel] = current_sum/num_fch[channel]
            channels_shap = channels_shap.append(pred_for_channnel, ignore_index=True)

        print ('shap for channel for lumi')
        print (channels_shap)
        print(" ")
        
        print ('shap for channel for run')
        print(channels_shap.sum())
        print(" ")
        print("_"*100)
        
    else:
        print (str(run) +' is not found')









    



276237 is not found
273426 is not found

(278821, ' [[1, 33], [36, 37]],  FED652 in error, EE+04 is off')

sum of features' influences from all lumis, 20 the most important

Index([u'qCCPhi5x5_0', u'qCCPhi5x5_5', u'qPFMetPhi_1', u'qCCEn5x5_5',
       u'qCCEta5x5_0', u'qCalJet1Pt_4', u'qEEenergy_3', u'qEEix_4',
       u'qESenergy_4', u'qEBchi2_1', u'qEEiy_0', u'qCalMETBEFOPhi_3',
       u'qCalMETPhi_3', u'qCalMETBEFOPhi_1', u'qPreShEn_5', u'qHBHEtime_5',
       u'qCalJet2Pt_5', u'qEBtime_3', u'qPreShEta_5', u'qPFJet8CHS1Eta_1'],
      dtype='object')

the most common features with big shap values

Index([u'qCCPhi5x5_0', u'qCCPhi5x5_5', u'qCCEn5x5_5', u'qCCEta5x5_0',
       u'qEEix_4', u'qESenergy_4', u'qEBchi2_1', u'qEEiy_0', u'qHBHEtime_5',
       u'qPreShEn_5', u'qCalJet2Pt_5', u'qPFJetEta_1', u'qPreShPhi_0',
       u'qCalJetEn_0', u'qPFJet8CHSEta_3', u'qPFJet8CHSPt_3'],
      dtype='object')

summary plot






    












    



reference plot for good lumis






    












    



shap for channel for lumi
          PF      calo     muons   photons
0  -0.000598 -0.001788  0.000524  0.000061
1  -0.000078 -0.001764 -0.000044 -0.000226
2  -0.000517 -0.002080  0.000039  0.000312
3  -0.000260 -0.002363  0.000691  0.000301
4  -0.000585 -0.001931  0.000054  0.000038
5  -0.000113 -0.006109  0.000233  0.000277
6  -0.000584  0.002204 -0.000035 -0.000130
7   0.000008 -0.002015 -0.000227  0.000229
8  -0.000351 -0.000415  0.000072  0.000366
9  -0.000200 -0.002013  0.000267  0.000294
10 -0.000684 -0.001810  0.000517  0.000195
11 -0.000290 -0.002200  0.000354  0.000184
12 -0.000073 -0.002373  0.000368  0.000039
13  0.000456 -0.002198  0.000727  0.000094
14 -0.000196 -0.000619  0.000036  0.000231
15  0.000031 -0.002056  0.000459  0.000125
16 -0.000449 -0.001790  0.000040  0.000088
17  0.000418 -0.001797  0.000665  0.000099
18  0.000406 -0.001966  0.000693  0.000129
19 -0.000077 -0.002025  0.000089  0.000020
20 -0.000580 -0.001687  0.000720 -0.000134
21 -0.000162 -0.001762  0.000953  0.000067
22 -0.000003 -0.002410  0.000365  0.000149
23 -0.000134 -0.002155  0.000144  0.000197
24 -0.000023 -0.001708  0.000624  0.000086
25 -0.000041 -0.002187  0.000783  0.000246
26 -0.000183 -0.002408  0.000126 -0.000171
27  0.000598 -0.002100  0.000305  0.000192
28 -0.000024 -0.002444  0.000878  0.000225
29  0.000589 -0.002332 -0.000074  0.000162
30  0.000340 -0.001825  0.000170  0.000152
31 -0.000525 -0.001655  0.000843  0.000137
32 -0.000620 -0.002100  0.000309  0.000242
 
shap for channel for run
PF        -0.004506
calo      -0.063880
muons      0.011669
photons    0.004276
dtype: float64
 
____________________________________________________________________________________________________
282408 is not found
282921 is not found

(275757, ' [[104, 122]],  low DCS: HBHE. nothing on PF plot')

sum of features' influences from all lumis, 20 the most important

Index([u'qPFJet4CHSEta_3', u'qEEchi2_3', u'qEEchi2_1', u'qPFJetEIEta_3',
       u'qEEchi2_0', u'qEBchi2_0', u'qHFiphi_1', u'qESenergy_3', u'qEEtime_3',
       u'qCCEta5x5_1', u'qPFMetPt_5', u'qPFMetPhi_4', u'qESenergy_4',
       u'qPreShEn_0', u'qESenergy_0', u'qCCEta5x5_5', u'qEBtime_3',
       u'qHBHEtime_3', u'qCalJet1Pt_0', u'qPFJetEta_3'],
      dtype='object')

the most common features with big shap values

Index([u'qPFJet4CHSEta_3', u'qEEchi2_3', u'qEEchi2_1', u'qPFJetEIEta_3',
       u'qESenergy_3', u'qEEchi2_0', u'qEBchi2_0', u'qHFiphi_1', u'qEEtime_3',
       u'qPFMetPt_5', u'qCCEta5x5_1', u'qPreShEn_0', u'qHBHEtime_3',
       u'qESenergy_4', u'qCCEta5x5_5', u'qESenergy_0', u'qEBtime_3',
       u'qSCEta5x5_4', u'qPreShYEta_5'],
      dtype='object')

summary plot






    












    



reference plot for good lumis






    












    



shap for channel for lumi
          PF      calo     muons   photons
0  -0.005485 -0.000995  0.000073 -0.000838
1  -0.004827  0.000368 -0.000181 -0.000726
2  -0.005052  0.000410 -0.000195 -0.000683
3  -0.004729 -0.000168  0.000102 -0.000124
4  -0.004679  0.000179 -0.000052 -0.000536
5  -0.005289  0.000420  0.000023 -0.000961
6  -0.004851  0.000445  0.000257 -0.000890
7  -0.004297  0.000198  0.000015 -0.000843
8  -0.004684  0.000245  0.000141 -0.000248
9  -0.004614  0.000361  0.000159 -0.000199
10 -0.005009 -0.000051  0.000283  0.000019
11 -0.004981  0.000356  0.000661 -0.000304
12 -0.005329  0.001099 -0.000074 -0.000666
13 -0.005405  0.000923  0.000374 -0.000823
14 -0.004767  0.000752  0.000353 -0.000311
15 -0.005052  0.001016  0.000379 -0.000259
16 -0.004907  0.000305  0.000403 -0.000431
 
shap for channel for run
PF        -0.083958
calo       0.005863
muons      0.002721
photons   -0.008822
dtype: float64
 
____________________________________________________________________________________________________

(275758, ' [[1, 4]],   part of hcal not available --> JetMET reconstruction suffers')

sum of features' influences from all lumis, 20 the most important

Index([u'qPFJet4CHSEta_3', u'qEEchi2_3', u'qEEchi2_1', u'qEEchi2_0',
       u'qEBchi2_0', u'qPFJetEIEta_3', u'qHFiphi_1', u'qESenergy_3',
       u'qEEtime_3', u'qCCEta5x5_1', u'qPFMetPhi_1', u'qESenergy_4',
       u'qESenergy_5', u'qESenergy_0', u'qPFJetEta_3', u'qCCEta5x5_5',
       u'qPFMetPhi_5', u'qEBtime_3', u'qMuCosmEta_1', u'qCalJet1Pt_0'],
      dtype='object')

the most common features with big shap values

Index([u'qPFJet4CHSEta_3', u'qEEchi2_3', u'qEEchi2_1', u'qEEchi2_0',
       u'qPFJetEIEta_3', u'qEBchi2_0', u'qESenergy_3', u'qEEtime_3',
       u'qESenergy_0', u'qCCEta5x5_1', u'qESenergy_4', u'qESenergy_5',
       u'qSCEta5x5_4', u'qPFMetPhi_5', u'qHBHEtime_3', u'qEEenergy_4',
       u'qEBtime_3', u'qCCEta5x5_5', u'qgedPhosigmaeta__4'],
      dtype='object')

summary plot






    












    



reference plot for good lumis






    












    



shap for channel for lumi
         PF      calo     muons   photons
0 -0.005070  0.000542  0.000347 -0.000184
1 -0.004338  0.000651 -0.000193 -0.000386
2 -0.004050  0.000890  0.000307 -0.000315
3 -0.000654 -0.000166  0.000237 -0.000834
 
shap for channel for run
PF        -0.014113
calo       0.001917
muons      0.000698
photons   -0.001719
dtype: float64
 
____________________________________________________________________________________________________

(275764, ' [[1, 31]],   pixel off')

sum of features' influences from all lumis, 20 the most important

Index([u'qEEchi2_3', u'qPFJet4CHSEta_3', u'qEEchi2_1', u'qPFMetPhi_5',
       u'qEEchi2_0', u'qEBchi2_0', u'qPFMetPhi_1', u'qPFJetEIEta_3',
       u'qESenergy_3', u'qEEtime_3', u'qHFiphi_1', u'qESenergy_4',
       u'qPFMetPhi_4', u'qCCEta5x5_1', u'qPFJetPhi_3', u'qPFMetPt_5',
       u'qPreShEn_0', u'qESenergy_0', u'qSCEta5x5_4', u'qEEenergy_4'],
      dtype='object')

the most common features with big shap values

Index([u'qEEchi2_3', u'qEEchi2_0', u'qEBchi2_0', u'qPFMetPhi_1',
       u'qESenergy_3', u'qEEtime_3', u'qESenergy_4'],
      dtype='object')

summary plot






    












    



reference plot for good lumis






    












    



shap for channel for lumi
          PF      calo     muons       photons
0  -0.000933  0.000063  0.000027  3.051029e-04
1   0.000066 -0.000358 -0.000046 -6.448529e-04
2  -0.001168 -0.000350 -0.000372 -4.304374e-04
3  -0.001169 -0.000736  0.000057  9.083104e-07
4  -0.001454 -0.000379 -0.000027 -3.198154e-04
5  -0.000447 -0.000631 -0.000479  1.447731e-04
6   0.000140 -0.000631  0.000148 -3.815831e-05
7  -0.002122  0.000586 -0.000105 -6.479834e-04
8  -0.002096 -0.000529  0.000055  4.236579e-05
9  -0.002714 -0.000204 -0.000092 -4.722076e-04
10 -0.002819 -0.000069  0.000224 -5.339483e-04
11 -0.004745  0.001382 -0.000091 -7.791587e-04
12 -0.004788  0.002204  0.000185  8.265715e-06
13 -0.004538  0.001567 -0.000045 -2.267468e-05
14 -0.004655  0.002098  0.000068  3.073530e-04
15 -0.004963  0.001697  0.000277 -1.002856e-03
16 -0.001904  0.000217 -0.000015  1.102260e-04
17 -0.005144  0.001132 -0.000348 -4.935323e-04
18 -0.005275  0.001595  0.000075 -8.728633e-05
19 -0.005011  0.001424 -0.000116 -2.740382e-04
20 -0.005264  0.001902  0.000091 -1.241774e-04
21 -0.004909  0.001158 -0.000400  1.523427e-04
22 -0.003464  0.000479 -0.000301 -7.463955e-04
23 -0.003045  0.000175 -0.000187  1.071306e-04
24 -0.002351 -0.000041 -0.000026 -9.652054e-04
25 -0.003006  0.000491 -0.000165 -7.409144e-04
26 -0.004673  0.001897 -0.000084 -4.541396e-04
27 -0.001499 -0.000033 -0.000328 -5.165425e-04
28 -0.003720  0.001489  0.000141 -6.294361e-04
29 -0.002275  0.000038 -0.000354 -6.625712e-04
 
shap for channel for run
PF        -0.089944
calo       0.017634
muons     -0.002233
photons   -0.009408
dtype: float64
 
____________________________________________________________________________________________________

(275766, ' [[1, 23], [25, 60]],  lower tracker efficiency')

sum of features' influences from all lumis, 20 the most important

Index([u'qEEchi2_3', u'qPFJet4CHSEta_3', u'qPFMetPhi_5', u'qEEchi2_1',
       u'qPFMetPhi_1', u'qEEchi2_0', u'qPFJetEIEta_3', u'qEBchi2_0',
       u'qESenergy_3', u'qEEtime_3', u'qPFMetPhi_4', u'qPFMetPt_5',
       u'qHFiphi_1', u'qESenergy_4', u'qESenergy_0', u'qPFJetPhi_3', u'lumi',
       u'qCCEta5x5_5', u'qPFMetPhi_0', u'qCCEta5x5_1'],
      dtype='object')

the most common features with big shap values

Index([u'qPFMetPhi_5', u'qEEchi2_1', u'qPFMetPhi_1', u'qESenergy_3',
       u'qEEtime_3', u'qPFMetPhi_4', u'qESenergy_0', u'qPFChMetPhi_0',
       u'qEBtime_3'],
      dtype='object')

summary plot






    












    



reference plot for good lumis






    












    



shap for channel for lumi
          PF      calo     muons   photons
0  -0.002241 -0.000250 -0.000305 -0.000284
1  -0.004957  0.001845 -0.000062 -0.000212
2  -0.001628 -0.000134  0.000039  0.000275
3  -0.002742  0.000448  0.000103 -0.000464
4  -0.004790  0.001885 -0.000149 -0.000766
5  -0.004990  0.001552 -0.000144 -0.000122
6  -0.002949  0.000326 -0.000104 -0.000273
7  -0.005196  0.001433 -0.000258 -0.000600
8  -0.002234 -0.000347 -0.000058 -0.000785
9  -0.002310  0.000142 -0.000016 -0.000537
10 -0.004540  0.001525 -0.000870 -0.000633
11 -0.002137 -0.000184  0.000264 -0.000609
12 -0.004297  0.001401 -0.000236 -0.000068
13 -0.004780  0.002001  0.000391 -0.000235
14 -0.004827  0.001942 -0.000375 -0.000629
15 -0.005120  0.001272 -0.000502 -0.000350
16 -0.005079  0.001732 -0.000131 -0.000344
17 -0.001993 -0.000362  0.000256 -0.000552
18 -0.004378  0.001622  0.000271  0.000472
19 -0.002106  0.000749 -0.000274 -0.000570
20 -0.004631  0.001654  0.000211 -0.000091
21 -0.001735  0.000590  0.000093 -0.000526
22 -0.004971  0.001800 -0.000310 -0.000429
23 -0.004536  0.001606 -0.000361 -0.000587
24 -0.004595  0.001941  0.000015 -0.000220
25 -0.001962 -0.000040 -0.000310 -0.000035
26 -0.002893  0.000680 -0.000104 -0.000153
27 -0.002620  0.000267  0.000189 -0.000587
28 -0.002070 -0.000081  0.000079 -0.000421
29 -0.002518 -0.000154 -0.000469 -0.000036
30 -0.002933  0.000475 -0.000282 -0.000831
31 -0.001556  0.000955 -0.000128 -0.000567
32 -0.001964  0.000067  0.000015  0.000366
33 -0.002030 -0.000397  0.000032  0.000519
34 -0.005197  0.001501 -0.000697 -0.000197
35 -0.004663  0.001785 -0.000291 -0.000296
36 -0.004383  0.002294 -0.000340  0.000209
37 -0.002461  0.000205 -0.000360 -0.000459
38 -0.002250  0.000058 -0.000140  0.000196
39 -0.001508  0.000020 -0.000231  0.000068
40 -0.001526 -0.000013  0.000149  0.000709
41 -0.001450 -0.000143  0.000113 -0.000222
42 -0.001641 -0.000052  0.000055  0.000513
43 -0.002002  0.000477 -0.000534  0.000174
44 -0.001518 -0.000047  0.000002  0.000624
45 -0.001579  0.000042 -0.000469 -0.000016
46 -0.001523  0.000176  0.000142  0.000468
47 -0.001663  0.000022 -0.000238 -0.000462
48 -0.001546  0.000476  0.000070  0.000268
49 -0.001217  0.000434  0.000217  0.000715
50 -0.001299  0.000842  0.000055  0.000633
51 -0.001417  0.000417  0.000330  0.000370
52 -0.001202  0.000399 -0.000163  0.000084
53 -0.001344 -0.000346 -0.000069 -0.001323
 
shap for channel for run
PF        -0.155694
calo       0.036509
muons     -0.005889
photons   -0.008829
dtype: float64
 
____________________________________________________________________________________________________

(275768, ' [[1, 79]],  lower tracker efficiency')

sum of features' influences from all lumis, 20 the most important

Index([u'qPFMetPhi_5', u'qEEchi2_1', u'qPFMetPhi_1', u'qESenergy_3',
       u'qPFMetPhi_4', u'qEEtime_3', u'qPFMetPhi_0', u'qPFMetPt_5',
       u'qPFJetPhi_3', u'lumi', u'qESenergy_0', u'qPFChMetPhi_1', u'qEBtime_3',
       u'qPFChMetPhi_0', u'qCalJet0En_4', u'qCalJet2Pt_5', u'qPFChMetPhi_4',
       u'qCalJet2Pt_0', u'qMuEta_5', u'qCCEta5x5_5'],
      dtype='object')

the most common features with big shap values

Index([u'qPFMetPhi_5', u'qEEchi2_1', u'qPFMetPhi_1', u'qPFMetPhi_4',
       u'qESenergy_3', u'qEEtime_3', u'qESenergy_0', u'qEBtime_3',
       u'qPFChMetPhi_4', u'qPFChMetPhi_0'],
      dtype='object')

summary plot






    












    



reference plot for good lumis






    












    



shap for channel for lumi
          PF      calo     muons   photons
0  -0.001563  0.000747  0.000043 -0.000287
1  -0.001432  0.000161 -0.000057  0.000275
2  -0.001121  0.000265  0.000182  0.000242
3  -0.001922 -0.000106  0.000240  0.000358
4  -0.001721  0.000124 -0.000077  0.000483
5  -0.002095  0.000018 -0.000065  0.000155
6  -0.001989  0.000528  0.000003  0.000507
7  -0.001493  0.000468  0.000119  0.000424
8  -0.001917  0.000820 -0.000022  0.000403
9  -0.001487  0.000928 -0.000327  0.000632
10 -0.001869  0.000158  0.000284  0.000413
11 -0.001947  0.000845  0.000116  0.000385
12 -0.001276 -0.000112 -0.000090 -0.000089
13 -0.000996  0.000271  0.000031  0.000375
14 -0.001346  0.000158  0.000042  0.000381
15 -0.001848  0.000664  0.000062  0.000500
16 -0.002127  0.000434  0.000075  0.000391
17 -0.001836 -0.000047  0.000170 -0.000514
18 -0.002299  0.000116  0.000055  0.000527
19 -0.000853 -0.000187 -0.000037  0.000146
20 -0.001787 -0.000569  0.000446  0.000213
21 -0.001848 -0.000109  0.000081  0.000666
22 -0.001513  0.000336  0.000159  0.000495
23 -0.001735  0.000509  0.000099  0.000140
24 -0.001632  0.000799  0.000223  0.000209
25 -0.001977  0.000005 -0.000111  0.000590
26 -0.001957 -0.000110 -0.000195  0.000815
27 -0.001711  0.000823 -0.000203  0.000246
28 -0.001968  0.000574 -0.000333  0.000487
29 -0.001747  0.000109 -0.000112  0.000365
..       ...       ...       ...       ...
46 -0.001558  0.000611 -0.000282  0.000216
47 -0.001497  0.001166  0.000420  0.000342
48 -0.001918  0.000571  0.000309  0.000551
49 -0.001460  0.000816 -0.000599  0.000476
50 -0.001638  0.000254  0.000210  0.000138
51 -0.000976  0.001002  0.000058  0.000245
52 -0.001375  0.000036 -0.000253  0.000378
53 -0.001528  0.001127 -0.000441  0.000547
54 -0.001584 -0.000061  0.000090  0.000522
55 -0.001965  0.000634  0.000058  0.000391
56 -0.000287  0.000728  0.000128  0.000422
57 -0.001489  0.000841  0.000054  0.000508
58 -0.001284  0.001085  0.000144  0.000662
59 -0.001713  0.000803  0.000146  0.000467
60 -0.001807  0.000716  0.000060  0.000234
61 -0.001757  0.000364 -0.000297 -0.000278
62 -0.001718  0.000886  0.000171 -0.000042
63 -0.001704  0.000702  0.000154  0.000399
64 -0.001706  0.000106  0.000022  0.000513
65 -0.001718  0.000098  0.000041  0.000267
66 -0.001414  0.000502  0.000017  0.000380
67 -0.002450  0.000623  0.000107  0.000194
68 -0.001710  0.000571 -0.000048  0.000384
69 -0.001597  0.000857  0.000229  0.000073
70 -0.001570  0.000155  0.000238  0.000115
71 -0.000490  0.000822 -0.000119  0.000161
72 -0.001738  0.000558  0.000439  0.000690
73 -0.001641  0.000346  0.000145  0.000078
74 -0.001514  0.000846  0.000284  0.000069
75 -0.001361  0.000307 -0.000152  0.000118

[76 rows x 4 columns]
 
shap for channel for run
PF        -0.123320
calo       0.034896
muons      0.002480
photons    0.022961
dtype: float64
 
____________________________________________________________________________________________________
275769 is not found
281663 is not found

(275781, ' [[1, 29]],  EB+16, +17, +18 are excluded due to a problem with VME crateS206h')

sum of features' influences from all lumis, 20 the most important

Index([u'qPFMetPhi_1', u'qEEtime_3', u'qEEenergy_3', u'qESenergy_3',
       u'qEBiPhi_1', u'qPFJetTopCHSPhi_3', u'qPFJet4CHS0Phi_4', u'qCCEta5x5_5',
       u'qCalMETBEFOPhi_4', u'qPho0En_5', u'qCalMETMPhi_4', u'qPFJet8CHSEta_3',
       u'qEEenergy_2', u'qPFJet3Pt_0', u'qPFJet8CHSPhi_0', u'qCalJet2Pt_0',
       u'qHBHEtime_3', u'qCalJet2Pt_5', u'qCalMETEn_0', u'qCalMETBEPhi_5'],
      dtype='object')

the most common features with big shap values

Index([u'qPFMetPhi_1', u'qCCEta5x5_5', u'qESenergy_3', u'qEBiPhi_1',
       u'qPFJet8CHSEta_3', u'qCalJet2Pt_0', u'qPFJet8CHSPt_3'],
      dtype='object')

summary plot






    












    



reference plot for good lumis






    












    



shap for channel for lumi
          PF      calo         muons   photons
0  -0.000801 -0.000925 -5.320194e-06 -0.000093
1  -0.001241 -0.000558  2.723344e-04 -0.000192
2  -0.001182 -0.000881  2.509181e-04  0.000075
3   0.000275  0.000419  8.722355e-05  0.000395
4   0.000224 -0.000441  2.805568e-04 -0.000048
5  -0.000366 -0.000087  2.406697e-04 -0.000342
6   0.000030 -0.001264  1.052117e-04  0.000109
7   0.000083 -0.000500  2.570417e-04 -0.000161
8  -0.000377  0.000062  1.098029e-04 -0.000010
9  -0.000009  0.000314 -2.597543e-05  0.000244
10  0.000155  0.000490 -1.477297e-04 -0.000217
11 -0.000179 -0.000284  3.344641e-04 -0.000156
12 -0.000022 -0.000984  8.619790e-05 -0.000288
13  0.000255 -0.000360  7.144160e-05  0.000345
14 -0.000665  0.000231  1.603744e-04 -0.000121
15  0.000391  0.000131  3.132582e-04  0.000087
16  0.000148 -0.000397  2.821776e-04  0.000080
17 -0.000256 -0.001014  2.486359e-07  0.000300
18  0.000119 -0.000130  1.698936e-04 -0.000156
19 -0.001090 -0.000480  2.576895e-04 -0.000141
20 -0.000628 -0.000327 -1.294332e-04 -0.000188
21  0.000604 -0.000236  3.355831e-04  0.000271
22  0.000337  0.000410  5.578710e-05  0.000217
23  0.000031  0.000329  1.539022e-04 -0.000101
24 -0.000029 -0.000997  3.784780e-04  0.000132
25  0.000088  0.000234  8.540020e-07  0.000014
26  0.000492 -0.000224  3.085468e-04  0.000021
27  0.000197 -0.000721  4.066144e-04  0.000147
28  0.000396 -0.000436  2.417866e-04  0.000246
 
shap for channel for run
PF        -0.003022
calo      -0.008627
muons      0.004853
photons    0.000470
dtype: float64
 
____________________________________________________________________________________________________
275783 is not found
281674 is not found
281680 is not found
282707 is not found
276064 is not found
276071 is not found
274282 is not found
281974 is not found
275838 is not found

(276095, ' [[1, 5]],   low stat')

sum of features' influences from all lumis, 20 the most important

Index([u'lumi', u'qPFJet4CHSEta_3', u'qEEchi2_1', u'qEBenergy_5',
       u'qPFJetEta_3', u'qEBchi2_0', u'qPFMetPt_5', u'qESenergy_3',
       u'qPFJetEIEta_3', u'qPFJet8CHS5Eta_0', u'qHFiphi_1', u'qPFJet8CHS5Pt_1',
       u'qHBHEtime_3', u'qEEenergy_3', u'qPFJet5Eta_0', u'qMu1Eta_2',
       u'qCalJet0En_1', u'qPhor1x5__3', u'qPreShEta_0',
       u'qgedPhomaxenxtal__1'],
      dtype='object')

the most common features with big shap values

Index([u'lumi', u'qEEchi2_1', u'qPFJetEta_3', u'qEBchi2_0', u'qPFMetPt_5',
       u'qESenergy_3', u'qEEenergy_3', u'qEBtime_4'],
      dtype='object')

summary plot






    












    



reference plot for good lumis






    












    



shap for channel for lumi
         PF      calo     muons   photons
0 -0.000902  0.000124 -0.000091 -0.000593
1 -0.001142  0.000294  0.000050 -0.000507
2 -0.001142 -0.000521 -0.000561 -0.000405
3 -0.000758  0.000363  0.000539  0.000316
4 -0.006053 -0.000776 -0.000465 -0.000748
 
shap for channel for run
PF        -0.009998
calo      -0.000516
muons     -0.000529
photons   -0.001938
dtype: float64
 
____________________________________________________________________________________________________
273296 is not found
282796 is not found
277933 is not found
280239 is not found
280241 is not found

(279995, ' [[1, 8]],  hcal water colling issues')

sum of features' influences from all lumis, 20 the most important

Index([u'qPFJet4CHSEta_3', u'qPFMetPt_5', u'qgedPhor2x5__1',
       u'qPFJet4CHSPhi_1', u'qEStime_1', u'qCalMETMEn_5', u'qPFMetPhi_3',
       u'qPFJetEIEta_3', u'qHFieta_1', u'qEEenergy_3', u'qPFJetEta_3',
       u'qNVtx_5', u'qPFJet4CHS3Phi_1', u'qEEenergy_5', u'qPFJet4CHS2Pt_2',
       u'qPFJet4CHSEta_5', u'qNVtx_3', u'qUNdrSumEt_0', u'qCalMETEn_5',
       u'qCalJet0En_3'],
      dtype='object')

the most common features with big shap values

Index([u'qPFJet4CHSEta_3', u'qPFMetPt_5', u'qgedPhor2x5__1',
       u'qPFJet4CHSPhi_1', u'qEStime_1', u'qCalMETMEn_5', u'qPFMetPhi_3',
       u'qPFJetEIEta_3', u'qHFieta_1', u'qEEenergy_3', u'qPFJetEta_3',
       u'qNVtx_5', u'qPFJet4CHS3Phi_1', u'qEEenergy_5', u'qPFJet4CHS2Pt_2',
       u'qPFJet4CHSEta_5', u'qNVtx_3', u'qUNdrSumEt_0', u'qCalMETEn_5',
       u'qCalJet0En_3', u'qUNSigmaIPhi_3', u'qPFJetEI0Phi_2', u'qMuCosmEta_1',
       u'qSCEn5x5_4', u'qCalMETEn_4', u'qPhosigmaeta__3', u'qCalJet3Pt_5',
       u'qPFJetPhi_1', u'qPFJetEta_0', u'qCalJet0En_1', u'qHBHEtime_0',
       u'qMuCosmLegEta_5', u'qESenergy_5', u'qHFiphi_1', u'qCalJet2Pt_5',
       u'qEBtime_3', u'qPFJet8CHS0Pt_4', u'qPho3Pt_1', u'qPFChMetPhi_0',
       u'qPFJetTopCHS5Phi_1', u'qCalMETEn_0', u'qPreShEn_5',
       u'qPFJet8CHSSD3Phi_5', u'qPhor1x5__3', u'qCalMETMPhi_3', u'qPhor9__3',
       u'qPFMetPhi_1', u'qPhor2x5__5', u'qPFMetPt_3', u'qCCEta5x5_5'],
      dtype='object')

summary plot






    












    



reference plot for good lumis






    












    



shap for channel for lumi
         PF      calo     muons  photons
0 -0.006691 -0.002853 -0.000428 -0.00099
 
shap for channel for run
PF        -0.006691
calo      -0.002853
muons     -0.000428
photons   -0.000990
dtype: float64
 
____________________________________________________________________________________________________
273445 is not found
280002 is not found

(280006, ' [[1, 68]],  EB-11 token ring missing')

sum of features' influences from all lumis, 20 the most important

Index([u'qCalMETBEFOPhi_1', u'qPFMetPhi_1', u'qCalMETPhi_3', u'qPFMetPhi_3',
       u'qPFMetPhi_5', u'qCalMETEn_4', u'qgedPhoPhi_1', u'qEBtime_0',
       u'qSCEn5x5_4', u'qESenergy_4', u'qPhosigmaeta__3', u'qPFJet8CHS0Eta_4',
       u'qPreShEn_5', u'qgedPho0Eta_5', u'qPFMetPt_0', u'qPFJet8CHSPt_1',
       u'qESenergy_5', u'qCCEta5x5_1', u'qPreShYEn_5', u'qPFMetPt_4'],
      dtype='object')

the most common features with big shap values

Index([u'qPFMetPhi_1', u'qPFMetPhi_3', u'qSCEn5x5_4', u'qPreShEn_5',
       u'qdrSumPt_4'],
      dtype='object')

summary plot






    












    



reference plot for good lumis






    












    



shap for channel for lumi
          PF          calo     muons       photons
0  -0.000028 -3.597893e-03  0.000659 -1.238716e-04
1  -0.000337 -4.303189e-03  0.000579  9.541683e-05
2  -0.000296  1.491182e-04  0.000002 -1.573623e-04
3  -0.000412 -4.608001e-03  0.000837 -2.715065e-04
4  -0.000631 -3.847228e-03  0.000565 -9.982572e-05
5  -0.000770 -5.690888e-03  0.000819 -2.505961e-05
6  -0.000664 -4.365352e-03  0.000757 -7.040532e-04
7  -0.000916 -5.270852e-03  0.000367 -1.301573e-04
8  -0.000240 -3.397774e-03  0.000842 -4.520137e-04
9  -0.000505 -5.161431e-03  0.000804  2.567742e-04
10 -0.000706 -4.835694e-03  0.000263 -6.476775e-04
11 -0.000676 -5.794688e-03  0.000453  1.559401e-04
12 -0.000775  2.713900e-04  0.000049 -3.251100e-04
13 -0.000230 -4.843772e-03  0.000716  2.038886e-04
14 -0.000851 -5.796085e-03 -0.000083 -1.219097e-04
15 -0.000754  1.709698e-04  0.000422 -1.157222e-04
16 -0.000596 -1.067964e-03 -0.000023 -3.643292e-04
17 -0.000798 -5.701038e-03  0.000537  9.246719e-07
18  0.000425  4.015918e-04  0.000296 -3.613942e-04
19 -0.000010 -8.109236e-04  0.000350 -3.298611e-04
20 -0.000516 -5.973122e-03  0.000440  1.605551e-05
21 -0.000364 -5.337275e-03  0.000624 -2.972993e-04
22 -0.000366  1.715153e-04  0.000207  1.468057e-04
23 -0.000399 -4.615123e-03  0.000756 -1.353706e-04
24 -0.000553 -9.971363e-04  0.000392 -2.900490e-04
25 -0.000369 -7.796324e-05  0.000195  2.789571e-04
26 -0.000609 -7.133368e-03  0.000910 -3.688056e-04
27 -0.000704 -5.857168e-03  0.000513  1.350762e-04
28 -0.000128  5.170970e-04  0.000323 -3.092312e-04
29 -0.000351 -2.271861e-04  0.000302 -5.164110e-04
..       ...           ...       ...           ...
36 -0.000494  5.076337e-04  0.000434  2.528402e-04
37 -0.000590 -4.828189e-04 -0.000285 -2.805516e-04
38 -0.000577  1.504082e-04  0.000138 -1.228704e-04
39 -0.000327 -3.345753e-03  0.000920 -1.515056e-04
40 -0.000871 -4.898873e-03  0.000493 -1.465604e-04
41 -0.000757 -1.075618e-03  0.000079 -1.449664e-04
42 -0.000130 -4.801745e-03  0.000630 -8.125157e-05
43 -0.000263 -2.839833e-03  0.000640 -2.168739e-04
44 -0.000508 -5.216862e-03  0.000738 -2.858890e-04
45 -0.000596  2.832418e-04  0.000262  4.657670e-05
46  0.000193  2.071445e-04  0.000134 -5.371417e-04
47 -0.000607 -5.107100e-03  0.000505 -5.190707e-04
48 -0.000061 -4.969106e-04  0.000257  1.260999e-04
49 -0.000515  4.845941e-05  0.000361  2.150197e-06
50 -0.000381 -5.265963e-03  0.000572 -1.106283e-04
51 -0.000457 -5.183117e-03  0.000745 -7.575110e-05
52 -0.000763 -7.895201e-04 -0.000280  6.955251e-05
53 -0.000226 -6.039066e-07 -0.000207 -5.350153e-04
54 -0.000529 -5.572547e-04  0.000211 -4.964396e-04
55 -0.000183 -3.949309e-04  0.000289 -4.105583e-04
56 -0.000111 -5.441553e-03  0.000855 -3.191125e-04
57 -0.000535 -8.594050e-04  0.000473 -6.100542e-05
58 -0.000446 -7.621454e-04  0.000276 -1.454007e-04
59 -0.000518 -3.287768e-03  0.000551  1.145696e-04
60 -0.000009  7.943087e-05 -0.000140 -2.366114e-04
61 -0.000978 -5.264135e-03 -0.000035  1.773473e-04
62 -0.000297  3.166200e-04  0.000303 -7.778931e-05
63 -0.000507 -5.536739e-03  0.000671 -3.870027e-04
64 -0.000726 -5.609164e-03  0.000701  7.203086e-05
65 -0.001009 -3.060686e-03  0.000542 -3.142829e-04

[66 rows x 4 columns]
 
shap for channel for run
PF        -0.029835
calo      -0.188992
muons      0.025960
photons   -0.009983
dtype: float64
 
____________________________________________________________________________________________________

(280007, ' [[1, 36]],  Low voltage channel broken in EB')

sum of features' influences from all lumis, 20 the most important

Index([u'qPFJet4CHSEta_3', u'qPFMetPt_5', u'qCalMETMEn_5', u'qPFJetEIEta_3',
       u'qCalMETEn_5', u'qHFieta_1', u'qPFJet8CHS0Eta_4', u'qEBenergy_5',
       u'qCalJet0Eta_0', u'qUNdrSumEt_0', u'qCalJet1Pt_0', u'qPFJet8CHS5Pt_1',
       u'qPFMetPhi_1', u'qPhosigmaeta__3', u'qSigmaIPhi_5', u'qPFJet8CHSEta_0',
       u'qeSCOP_6', u'qHBHEtime_0', u'qEBtime_3', u'qPFJetEta_3'],
      dtype='object')

the most common features with big shap values

Index([u'qPFJet4CHSEta_3', u'qPFMetPt_5', u'qCalMETMEn_5', u'qPFJetEIEta_3',
       u'qCalMETEn_5', u'qHFieta_1', u'qPFJet8CHS0Eta_4', u'qCalJet0Eta_0',
       u'qCalJet1Pt_0', u'qPFJet8CHSEta_0', u'qPhosigmaeta__3', u'qEBtime_3',
       u'qHBHEtime_0', u'qSigmaIPhi_5', u'qSCEtaWidth_3', u'qEBchi2_0',
       u'qCalMETEn_4', u'qSCEn5x5_4', u'qPreShYEn_3', u'qEEtime_5'],
      dtype='object')

summary plot






    












    



reference plot for good lumis






    












    



shap for channel for lumi
         PF      calo     muons   photons
0 -0.007196 -0.002203  0.000309 -0.000739
1 -0.007978 -0.002012  0.000198 -0.000635
2 -0.006090 -0.002896  0.000341 -0.001039
 
shap for channel for run
PF        -0.021264
calo      -0.007111
muons      0.000848
photons   -0.002413
dtype: float64
 
____________________________________________________________________________________________________

(275922, ' [[4, 6]],  low stat')

sum of features' influences from all lumis, 20 the most important

Index([u'qEEtime_3', u'qEBtime_4', u'qHFiphi_1', u'qESenergy_4',
       u'qESenergy_3', u'qEEenergy_3', u'qPreShYEn_4', u'qHBHEauxe_1',
       u'qPreShYEn_5', u'qCCEta5x5_5', u'qPFJet8CHSPt_4', u'lumi',
       u'qEEchi2_0', u'qHBHEtime_0', u'qPreShEn_4', u'qPFJetEI2Eta_1',
       u'qPreShEn_5', u'qCalJet2Phi_2', u'qHFieta_1', u'qPFJet8CHSSD4Eta_0'],
      dtype='object')

the most common features with big shap values

Index([u'qEBtime_4', u'qEEtime_3', u'qHFiphi_1', u'qESenergy_4',
       u'qEEenergy_3', u'qPreShYEn_4', u'qHBHEauxe_1', u'qPreShYEn_5',
       u'qEEchi2_0', u'qHBHEtime_0', u'qPreShEn_5', u'qCalJet2Phi_2',
       u'qHBHEtime_5', u'qMuEta_1', u'qPFJet8CHSEta_3', u'qHBHEauxe_5',
       u'qgedPhoe1x5__4', u'qPFJet8CHSPt_3'],
      dtype='object')

summary plot






    












    



reference plot for good lumis






    












    



shap for channel for lumi
         PF      calo     muons   photons
0  0.000182 -0.000348  0.000112 -0.000188
1 -0.000381 -0.000029  0.000525  0.000043
2 -0.000219  0.000014  0.000624 -0.000134
 
shap for channel for run
PF        -0.000418
calo      -0.000363
muons      0.001261
photons   -0.000279
dtype: float64
 
____________________________________________________________________________________________________
274142 is not found

(278309, ' [[1, 10]],  EE-04 FED 607 TT disabled in LS [1-10] according to express dataset (LS# 10)')

sum of features' influences from all lumis, 20 the most important

Index([u'qCCPhi5x5_0', u'qCCPhi5x5_5', u'qCCEta5x5_4', u'qCCEta5x5_0',
       u'qCCEn5x5_5', u'qCalJet1Pt_4', u'qPFMetPhi_1', u'qEEenergy_3',
       u'qCCEta5x5_5', u'qEEix_4', u'qESenergy_4', u'qHBHEtime_0',
       u'qEBtime_3', u'qEEiy_0', u'qPFJet4CHSEta_5', u'qPreShEn_4',
       u'qSCEta5x5_4', u'qPFJet8CHS1Eta_1', u'qPreShEn_5',
       u'qCalMETBEFOPhi_3'],
      dtype='object')

the most common features with big shap values

Index([u'qCCPhi5x5_0', u'qCCPhi5x5_5', u'qCCEta5x5_0', u'qCCEta5x5_4',
       u'qCCEn5x5_5', u'qPFMetPhi_1', u'qCalJet1Pt_4', u'qEEenergy_3',
       u'qCCEta5x5_5', u'qEEix_4', u'qESenergy_4', u'qEBtime_3',
       u'qPFJet4CHSEta_5', u'qPreShEn_4', u'qPreShEn_5', u'qHBHEtime_5',
       u'qHBHEtime_3', u'qPreShPhi_0', u'qCalJet2Pt_5', u'qPFJet8CHSEta_3',
       u'qCalJetEn_0'],
      dtype='object')

summary plot






    












    



reference plot for good lumis






    












    



shap for channel for lumi
         PF      calo     muons   photons
0  0.000019 -0.001586  0.000476  0.000121
1 -0.000576 -0.002041  0.000353 -0.000117
2  0.000034 -0.001599  0.000346  0.000016
3  0.000130 -0.002460  0.000218  0.000103
4 -0.000561 -0.002080  0.000200 -0.000081
5 -0.000358 -0.001944 -0.000324  0.000008
6  0.000044 -0.001894  0.000677 -0.000034
7 -0.000164 -0.002007  0.000440 -0.000081
 
shap for channel for run
PF        -0.001430
calo      -0.015612
muons      0.002386
photons   -0.000065
dtype: float64
 
____________________________________________________________________________________________________
277217 is not found

(276453, ' [[1, 8], [10, 125]],  EB-17 (FED 626): was excluded (because of cooling failure)')

sum of features' influences from all lumis, 20 the most important

Index([u'qSCEtaWidth_0', u'qEBiPhi_1', u'qEEenergy_3', u'qEBenergy_5',
       u'qgedPhoEta_1', u'qPFMetPhi_1', u'qSCPhiWidth_5', u'qHBHEtime_5',
       u'qPFJet8CHSEta_3', u'qCalJet2Pt_5', u'qHBHEtime_0',
       u'qgedPhosigmaeta__0', u'qUNSigmaIPhi_3', u'qCalJet2Pt_0',
       u'qHBHEauxe_5', u'qEEix_0', u'qCalJetEn_0', u'qPFJet8CHSPt_4',
       u'qPho0Pt_4', u'qHBHEauxe_1'],
      dtype='object')

the most common features with big shap values

Index([u'qEBiPhi_1', u'qPFJet8CHSEta_3'], dtype='object')

summary plot






    












    



reference plot for good lumis






    












    



shap for channel for lumi
           PF      calo     muons   photons
0    0.000306 -0.000211  0.000008 -0.000259
1   -0.000059 -0.000088  0.000178  0.000346
2    0.000374 -0.000120  0.000129  0.000265
3    0.000590 -0.000097  0.000111 -0.000039
4    0.000330  0.000163  0.000086 -0.000449
5    0.000353 -0.000313  0.000273 -0.000327
6    0.000431 -0.000186  0.000713  0.000152
7   -0.000596 -0.000257  0.000907 -0.001035
8   -0.000181 -0.000618  0.000705 -0.000335
9   -0.000075 -0.000490 -0.000173 -0.000416
10  -0.000025  0.000179  0.000351 -0.000457
11   0.000115  0.000140  0.000332 -0.000722
12  -0.000188 -0.000147  0.000116 -0.000501
13  -0.000287  0.000297  0.000659 -0.000946
14  -0.000131 -0.000562  0.000715 -0.000780
15   0.000050 -0.000387  0.000012 -0.000881
16  -0.000543 -0.000305  0.000415 -0.000387
17  -0.000209  0.000158  0.000210 -0.000670
18   0.000143  0.000059 -0.000184 -0.000592
19  -0.000115  0.000324  0.000660 -0.000829
20   0.000347  0.000049  0.000856 -0.000950
21   0.000310  0.000345  0.000065 -0.000977
22   0.000312 -0.000271  0.000116 -0.000820
23  -0.000074 -0.000566  0.000345 -0.000502
24  -0.000218 -0.000220  0.000641 -0.000314
25  -0.000025 -0.000404 -0.000009 -0.000125
26   0.000094  0.000053  0.000038  0.000038
27   0.000096 -0.000367  0.000069 -0.000545
28  -0.000466  0.000018  0.000015 -0.000584
29   0.000213 -0.000255  0.000963 -0.000084
..        ...       ...       ...       ...
77   0.000322 -0.000238  0.000339  0.000185
78  -0.000362 -0.000082  0.000141  0.000016
79   0.000336 -0.000303  0.000267  0.000301
80   0.000446 -0.000706  0.000236  0.000033
81   0.000267  0.000034  0.000047  0.000329
82   0.000706 -0.000039  0.000069 -0.000002
83   0.000555 -0.000026 -0.000087  0.000197
84   0.000596 -0.000172  0.000630  0.000172
85   0.000425 -0.000535  0.000052  0.000515
86   0.000617  0.000128  0.000109  0.000265
87   0.000413 -0.000791 -0.000048  0.000221
88  -0.000293 -0.000667 -0.000165 -0.000125
89   0.000427 -0.000131  0.000313  0.000332
90   0.000500 -0.000011  0.000171  0.000289
91  -0.000180 -0.000142  0.000530 -0.000064
92   0.000080  0.000023  0.000263  0.000276
93   0.000486 -0.000007  0.000211  0.000231
94  -0.000325  0.000058  0.000118 -0.000072
95   0.000590 -0.000216  0.000147  0.000232
96  -0.000200 -0.000065  0.000144 -0.000213
97  -0.000290 -0.000159  0.000093 -0.000340
98   0.000457 -0.000136  0.000300 -0.000093
99   0.000552 -0.000082  0.000174  0.000208
100  0.000797 -0.000469  0.000127  0.000180
101 -0.000086 -0.000033  0.000158 -0.000358
102 -0.000370  0.000044 -0.000017 -0.000037
103  0.000555  0.000051  0.000648  0.000091
104  0.000374  0.000009  0.000098  0.000214
105  0.000629 -0.000475  0.000187  0.000085
106  0.000336 -0.000884  0.000033 -0.000016

[107 rows x 4 columns]
 
shap for channel for run
PF         0.022130
calo      -0.016740
muons      0.024457
photons   -0.004883
dtype: float64
 
____________________________________________________________________________________________________
276455 is not found
276456 is not found
276457 is not found
274157 is not found
279028 is not found
275326 is not found
273150 is not found



In [64]:

    
sublabels[sublabels["runId"]==279995]









    Out[64]:






  
    
      
      runId
      lumiId
      isSig
      Csc
      Dt
      Ecal
      Egamma
      Es
      Hcal
      Hlt
      Jetmet
      L1tcalo
      L1tmu
      Lumi
      Muon
      Pix
      Rpc
      Strip
      Track
      new_json
    
  
  
    
      108963
      279995
      8
      0
      1
      1
      1
      1
      1
      0
      1
      0
      1
      1
      1
      1
      1
      1
      1
      1
      0.0



In [ ]: