WillML/MachineLearning.py at master · OceanNuclear/WillML · 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
import numpy as np
import pandas as pd
import pickle

UNLABELLED = -1
LOGSVM = False

#cd into the directory to be operated on.
try:
    import os, sys
    os.chdir(sys.argv[1])
    if len(sys.argv[2:])>0:
        exec(''.join(sys.argv[2:]))
except IndexError:
    print("usage:")
    print("'python "+sys.argv[0]+" <folder_containing_preprocessed_data>/'")
    exit()
#Read the training, testing, etc. data.
try:
    trfea = pd.read_pickle("trfea.pkl")
    trlab = pd.read_pickle("trlab.pkl")
    tefea = pd.read_pickle("tefea.pkl")
    telab = pd.read_pickle("telab.pkl")
except FileNotFoundError:
    print("Please cut up the data with 'Embed.py' first.")
    exit()

from sklearn.linear_model import LogisticRegression
from sklearn.ensemble import RandomForestClassifier
from sklearn.svm import LinearSVC
from sklearn.naive_bayes import MultinomialNB

from sklearn.model_selection import cross_val_score

random_forest = RandomForestClassifier(n_estimators=1000, max_depth=15) #max_depth=5 #yields the same result
svm           = LinearSVC()
naive_bayes   = MultinomialNB()
logreg        = LogisticRegression(random_state = 0, multi_class='auto', solver='newton-cg')

# random_forest.fit(trfea, trlab)
# print("random_forest obtained a score of", random_forest.score(tefea, telab))
# #(n_estimators=200, max_depth=5)  -> 0.4595119245701608
# #(n_estimators=200, max_depth=10) -> 0.704104
# #(n_estimators=200, max_depth=12) -> 0.7444536882972823
# #(n_estimators=1000, max_depth=15)-> 0.7827232390460344

# naive_bayes.fit(trfea, trlab)
# print("naive_bayes obtained a score of", naive_bayes.score(tefea, telab))

svm.fit(trfea, trlab)
#remove the unlablled telab cases
missing_entries = telab==UNLABELLED

if sum(~missing_entries) >0:
    print("linear svm (SVC) obtained an accuracy of {0} on the test set.".format(svm.score(tefea[~missing_entries], telab[~missing_entries])))
if LOGSVM:
    logreg.fit(trfea, trlab)
    if sum(~missing_entries) >0:
        print("logistic regression obtained an accuracy of {0} on the test set.".format(logreg.score(tefea[~missing_entries], telab[~missing_entries])))

linsvm_pred = svm.decision_function(tefea)
linsvm_pred = pd.DataFrame(linsvm_pred, index=telab.index)
if LOGSVM:
    logsvm_pred = logreg.decision_function(tefea)
    logsvm_pred = pd.DataFrame(logsvm_pred, index=telab.index)

with open('linsvm_pred.pkl', 'wb') as f:
    pickle.dump(linsvm_pred, f)
if LOGSVM:
    with open('logsvm_pred.pkl', 'wb') as f:
        pickle.dump(logsvm_pred, f)
#lbfgs      ->
#newton-cg  -> 0.947310038824182

'''
import tensorflow as tf
from tensorflow import keras
from tensorflow.keras import layers, activations
tf.random.set_random_seed(0)

# neural_network_structure = []
hyperparamteres             ={
                             'activaiton function' : activations.softmax,
                             'num_epochs': 10,
                             'learning_rate' :0.0001,
                             # 'loss_func' : 'binary_crossentropy', #tf.keras.losses.
                             'loss_func' : 'categorical_crossentropy',
                             'metrics': ['Accuracy'],
                             'regularize': True} #tf.keras.metrics.
print(trfea)
# mlresult = convert2onehot(sheet['label'])
# mlresult.to_pickle("mlresult.pkl")
'''