GlobalMuonTracking/python_training.py at master · Lucas-NL/GlobalMuonTracking · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
#!/usr/bin/python
import os
import sys
import math
import pickle
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split

import uproot
import lightgbm as lgb

from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.datasets import make_moons, make_circles, make_classification
from sklearn.neural_network import MLPClassifier
from sklearn.neighbors import KNeighborsClassifier
from sklearn.svm import SVC
from sklearn.gaussian_process import GaussianProcessClassifier
from sklearn.gaussian_process.kernels import RBF
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier, AdaBoostClassifier, GradientBoostingClassifier
from sklearn.naive_bayes import GaussianNB
from sklearn.discriminant_analysis import QuadraticDiscriminantAnalysis

names = ["Nearest Neighbors", "Linear SVM", "RBF SVM", "Gaussian Process",
         "Decision Tree", "Random Forest", "Neural Net", "AdaBoost",
         "Naive Bayes", "QDA","LightGBM"]

classifiers = [
    KNeighborsClassifier(3),
    SVC(kernel="linear", C=0.025),
    SVC(gamma=2, C=1),
    GaussianProcessClassifier(1.0 * RBF(1.0)),
    DecisionTreeClassifier(max_depth=5),
    RandomForestClassifier(max_depth=5, n_estimators=10, max_features=1),
    MLPClassifier(alpha=1, max_iter=1000),
    AdaBoostClassifier(),
    GradientBoostingClassifier(n_estimators=1000, learning_rate=0.01,max_depth=10, random_state=0),
    GaussianNB(),
    QuadraticDiscriminantAnalysis(),
    lgb
]

classifiers =[lgb]

file = uproot.open(os.environ['ML_TRAINING_FILE'])
matchTree = file["matchTree"]

MFT_X      = matchTree.array("MFT_X");
MFT_Y      = matchTree.array("MFT_Y");
MFT_Phi    = matchTree.array("MFT_Phi");
MFT_Tanl   = matchTree.array("MFT_Tanl");
MFT_InvQPt = matchTree.array("MFT_InvQPt");
MFT_Cov00  = matchTree.array("MFT_Cov00");
MFT_Cov01  = matchTree.array("MFT_Cov01");
MFT_Cov11  = matchTree.array("MFT_Cov11");
MFT_Cov02  = matchTree.array("MFT_Cov02");
MFT_Cov12  = matchTree.array("MFT_Cov12");
MFT_Cov22  = matchTree.array("MFT_Cov22");
MFT_Cov03  = matchTree.array("MFT_Cov03");
MFT_Cov13  = matchTree.array("MFT_Cov13");
MFT_Cov23  = matchTree.array("MFT_Cov23");
MFT_Cov33  = matchTree.array("MFT_Cov33");
MFT_Cov04  = matchTree.array("MFT_Cov04");
MFT_Cov14  = matchTree.array("MFT_Cov14");
MFT_Cov24  = matchTree.array("MFT_Cov24");
MFT_Cov34  = matchTree.array("MFT_Cov34");
MFT_Cov44  = matchTree.array("MFT_Cov44");
MCH_X      = matchTree.array("MCH_X");
MCH_Y      = matchTree.array("MCH_Y");
MCH_Phi    = matchTree.array("MCH_Phi");
MCH_Tanl   = matchTree.array("MCH_Tanl");
MCH_InvQPt = matchTree.array("MCH_InvQPt");
MCH_Cov00  = matchTree.array("MCH_Cov00");
MCH_Cov01  = matchTree.array("MCH_Cov01");
MCH_Cov11  = matchTree.array("MCH_Cov11");
MCH_Cov02  = matchTree.array("MCH_Cov02");
MCH_Cov12  = matchTree.array("MCH_Cov12");
MCH_Cov22  = matchTree.array("MCH_Cov22");
MCH_Cov03  = matchTree.array("MCH_Cov03");
MCH_Cov13  = matchTree.array("MCH_Cov13");
MCH_Cov23  = matchTree.array("MCH_Cov23");
MCH_Cov33  = matchTree.array("MCH_Cov33");
MCH_Cov04  = matchTree.array("MCH_Cov04");
MCH_Cov14  = matchTree.array("MCH_Cov14");
MCH_Cov24  = matchTree.array("MCH_Cov24");
MCH_Cov34  = matchTree.array("MCH_Cov34");
MCH_Cov44  = matchTree.array("MCH_Cov44");

MFT_TrackChi2   = matchTree.array("MFT_TrackChi2");
MFT_NClust      = matchTree.array("MFT_NClust");
CorrectMatching = matchTree.array("Truth");

Delta_X      = MCH_X      - MFT_X;
Delta_Y      = MCH_Y      - MFT_Y;
Delta_XY     = np.sqrt((MCH_X-MFT_X)**2 + (MCH_Y-MFT_Y)**2)
Delta_Phi    = MCH_Phi    - MFT_Phi;
Delta_Tanl   = MCH_Tanl   - MFT_Tanl;
Delta_InvQPt = MCH_InvQPt - MFT_InvQPt;
Delta_Cov00  = MCH_Cov00 - MFT_Cov00;
Delta_Cov01  = MCH_Cov01 - MFT_Cov01;
Delta_Cov11  = MCH_Cov11 - MFT_Cov11;
Delta_Cov02  = MCH_Cov02 - MFT_Cov02;
Delta_Cov12  = MCH_Cov12 - MFT_Cov12;
Delta_Cov22  = MCH_Cov22 - MFT_Cov22;
Delta_Cov03  = MCH_Cov03 - MFT_Cov03;
Delta_Cov13  = MCH_Cov13 - MFT_Cov13;
Delta_Cov23  = MCH_Cov23 - MFT_Cov23;
Delta_Cov33  = MCH_Cov33 - MFT_Cov33;
Delta_Cov04  = MCH_Cov04 - MFT_Cov04;
Delta_Cov14  = MCH_Cov14 - MFT_Cov14;
Delta_Cov24  = MCH_Cov24 - MFT_Cov24;
Delta_Cov34  = MCH_Cov34 - MFT_Cov34;
Delta_Cov44  = MCH_Cov44 - MFT_Cov44;

Ratio_X      = MCH_X      - MFT_X;
Ratio_Y      = MCH_Y      - MFT_Y;
Ratio_XY     = np.sqrt((MCH_X-MFT_X)**2 + (MCH_Y-MFT_Y)**2)
Ratio_Phi    = MCH_Phi    - MFT_Phi;
Ratio_Tanl   = MCH_Tanl   - MFT_Tanl;
Ratio_InvQPt = MCH_InvQPt - MFT_InvQPt;
Ratio_Cov00  = MCH_Cov00 - MFT_Cov00;
Ratio_Cov01  = MCH_Cov01 - MFT_Cov01;
Ratio_Cov11  = MCH_Cov11 - MFT_Cov11;
Ratio_Cov02  = MCH_Cov02 - MFT_Cov02;
Ratio_Cov12  = MCH_Cov12 - MFT_Cov12;
Ratio_Cov22  = MCH_Cov22 - MFT_Cov22;
Ratio_Cov03  = MCH_Cov03 - MFT_Cov03;
Ratio_Cov13  = MCH_Cov13 - MFT_Cov13;
Ratio_Cov23  = MCH_Cov23 - MFT_Cov23;
Ratio_Cov33  = MCH_Cov33 - MFT_Cov33;
Ratio_Cov04  = MCH_Cov04 - MFT_Cov04;
Ratio_Cov14  = MCH_Cov14 - MFT_Cov14;
Ratio_Cov24  = MCH_Cov24 - MFT_Cov24;
Ratio_Cov34  = MCH_Cov34 - MFT_Cov34;
Ratio_Cov44  = MCH_Cov44 - MFT_Cov44;

MFT_TrackReducedChi2 = MFT_TrackChi2/MFT_NClust;

training_list = np.stack([MFT_X,MFT_Y,MFT_Phi,MFT_Tanl,MFT_InvQPt,MFT_Cov00,MFT_Cov01,MFT_Cov11,MFT_Cov02,MFT_Cov12,MFT_Cov22,MFT_Cov03,MFT_Cov13,MFT_Cov23,MFT_Cov33,MFT_Cov04,MFT_Cov14,MFT_Cov24,MFT_Cov34,MFT_Cov44,
                          MCH_X,MCH_Y,MCH_Phi,MCH_Tanl,MCH_InvQPt,MCH_Cov00,MCH_Cov01,MCH_Cov11,MCH_Cov02,MCH_Cov12,MCH_Cov22,MCH_Cov03,MCH_Cov13,MCH_Cov23,MCH_Cov33,MCH_Cov04,MCH_Cov14,MCH_Cov24,MCH_Cov34,MCH_Cov44,
                          MFT_TrackChi2,MFT_NClust,MFT_TrackReducedChi2,
                          Delta_X,Delta_Y,Delta_XY,Delta_Phi,Delta_Tanl,Delta_InvQPt,Delta_Cov00,Delta_Cov01,Delta_Cov11,Delta_Cov02,Delta_Cov12,Delta_Cov22,Delta_Cov03,Delta_Cov13,Delta_Cov23,Delta_Cov33,Delta_Cov04,Delta_Cov14,Delta_Cov24,Delta_Cov34,Delta_Cov44,
                          Ratio_X,Ratio_Y,Ratio_XY,Ratio_Phi,Ratio_Tanl,Ratio_InvQPt,Ratio_Cov00,Ratio_Cov01,Ratio_Cov11,Ratio_Cov02,Ratio_Cov12,Ratio_Cov22,Ratio_Cov03,Ratio_Cov13,Ratio_Cov23,Ratio_Cov33,Ratio_Cov04,Ratio_Cov14,Ratio_Cov24,Ratio_Cov34,Ratio_Cov44], 1);

X = training_list
y = CorrectMatching

X = StandardScaler().fit_transform(X)

X_train, X_eval, y_train, y_eval = train_test_split(X, y,test_size=0.2,random_state=0,stratify=y)

evals_result = {}

train_data = lgb.Dataset(X_train, label=y_train)
valid_data = lgb.Dataset(X_eval, label=y_eval, reference = train_data)

params = {
    'task':'train',
    'boosting_type':'gbdt',
    'objective': 'binary',
    'learning_rate':0.01,
    'max_depth':10,
    'n_estimators':1000,
    'metric':'auc',
    'verbose': 2,
}

model = lgb.train(
    params,
    train_data,
    valid_sets=[train_data,valid_data],
    valid_names = ['train','eval'],
    num_boost_round=1000,
    early_stopping_rounds=100,
    evals_result=evals_result,
)

train_metric = evals_result['train']['auc']
eval_metric = evals_result['eval']['auc']
plt.plot(train_metric, label='train auc')
plt.plot(eval_metric, label='eval auc')
plt.grid()
plt.legend()
plt.ylim(0, 1.1)

plt.xlabel('rounds')
plt.ylabel('auc')
plt.show()