-
Notifications
You must be signed in to change notification settings - Fork 0
/
ML_PTP.py
37 lines (30 loc) · 1.22 KB
/
ML_PTP.py
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
import datetime
from MuDataFrame import MuDataFrame
from NN import TrainNN, PredictNN
from sklearn.model_selection import train_test_split
import datetime
import sys, os
#loading the input data
path2csvq = "/lustre/work/sshanto/PhotonTimeML/data_sets/calibration_new_set_up/calibration_new_set_up.csv"
path2csvl = "/Volumes/mac_extended/Research/MT/proto1b/data_sets/calibration_new_set_up/calibration_new_set_up.csv"
mdfo_calib = MuDataFrame(path2csvl)
mdf_calib = mdfo_calib.events_df
mdfo_calib.keep4by4Events()
all_df = mdfo_calib.events_df
all_df.drop(columns=[
'Unnamed: 0', 'event_time', 'Run_Num', 'time_of_day', 'SmallCounter',
'speed', 'xx', 'yy'
],
inplace=True)
df_train_all, df_test_all = train_test_split(all_df, test_size=0.3)
targetList = ["diffL2", "diffL4"]
inputList = [
'event_num', 'L1', 'R1', 'L3', 'R3', 'TopCounter', 'BottomCounter',
'diffL1', 'diffL3', 'sumL1', 'sumL3'
]
date = datetime.datetime.today().strftime('%Y-%m-%d-%H_%M_%S')
emissions = "Run_{}".format(date)
df_train_all.to_csv("train.csv")
df_test_all.to_csv("test.csv")
# TrainNN(df_train_all, df_test_all, targetList, inputList, emissions, epoch=1)
# PredictNN(df_test_all, df_train_all, targetList, inputList, emissions)