-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathklaas.yaml
175 lines (153 loc) · 3.53 KB
/
klaas.yaml
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
seed: 0
n_cross_validations: 5
selection:
num_pixel_in_shower:
- '>=' # needs to be in quotes because > and ! are special character in yaml
- 10
num_islands:
- <
- 8
length:
- <
- 70
width:
- <
- 35
leakage1:
- <
- 0.6
leakage2:
- < # string representation also possible
- 0.85
energy:
regressor: |
ensemble.RandomForestRegressor(
n_estimators=200,
max_features=5,
min_samples_split=3,
n_jobs=-1,
criterion='mse',
max_depth=15,
random_state=0,
)
n_signal: 120000
log_target: False
target_name: corsika_event_header_total_energy
features:
- size
- width
- length
- skewness_trans
- skewness_long
- concentration_cog
- concentration_core
- concentration_one_pixel
- concentration_two_pixel
- leakage1
- leakage2
- num_islands
- num_pixel_in_shower
- photoncharge_shower_mean
- photoncharge_shower_variance
- photoncharge_shower_max
# feature generation, constants have to be prefixed with @
feature_generation:
needed_columns:
- size
- width
- length
- cog_x
- cog_y
features:
log_size: log(size)
size_area: size / (width * length * @pi)
area: (width * length * @pi)
cog_r: sqrt(cog_x**2 + cog_y**2)
separator:
# the classifier to use
classifier : |
ensemble.RandomForestClassifier(
n_estimators=200,
max_features='sqrt',
n_jobs=-1,
max_depth=15,
criterion='entropy',
)
# randomly sample the data if you dont want to use the whole set
n_background: 120000
n_signal: 120000
features:
- concentration_cog
- concentration_core
- concentration_one_pixel
- concentration_two_pixel
- leakage1
- leakage2
- size
- width
- length
- skewness_long
- skewness_trans
- kurtosis_long
- kurtosis_trans
- num_islands
- num_pixel_in_shower
- photoncharge_shower_mean
- photoncharge_shower_variance
# feature generation, constants have to be prefixed with @
feature_generation:
needed_columns:
- width
- length
- size
features:
area: width * length * @pi
log_size: log(size)
log_length: log(length)
size_area: size / (width * length * @pi)
area_size_cut_var: (width * length * @pi) / log(size)**2
disp:
disp_regressor : |
ensemble.RandomForestRegressor(
n_estimators=200,
max_features='sqrt',
n_jobs=-1,
max_depth=20,
)
sign_classifier: |
ensemble.RandomForestClassifier(
n_estimators=200,
max_features='sqrt',
n_jobs=-1,
max_depth=20,
)
source_az_column: source_position_az
source_zd_column: source_position_zd
pointing_az_column: pointing_position_az
pointing_zd_column: pointing_position_zd
# randomly sample the data if you dont want to use the whole set
n_signal : 200000
features:
- num_pixel_in_shower
- width
- length
- skewness_long
- kurtosis_long
- concentration_cog
- concentration_core
- leakage1
- leakage2
- slope_long
- time_gradient_slope_long
- photoncharge_shower_mean
- photoncharge_shower_variance
feature_generation:
needed_columns:
- width
- length
- size
features:
area: width * length * @pi
width_length: 1 - (width / length)
log_size: log(size)
log_size_area: log(size) / (width * length * @pi)