USC-CS544/nbclassify.py at master · lpyeh/USC-CS544 · 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
#!usr/bin/python
import sys
import math
import os
import numpy as np
from collections import defaultdict
from collections import Counter
import glob
import re
import string
from preprocess import *

def load_data():
    all_files = glob.glob(os.path.join(input_path, '*/*/*/*.txt'))
    test = []
    paths = []
    for file_ in all_files:
        class1, class2, fold, fname = file_.split('/')[-4:]
        text = open(file_).read()
        test.append(text)
        paths.append(file_)
    test = clean(test)
    return test, paths


def load_params():
    log_priors = defaultdict(float)
    loglikelihood = defaultdict(dict)
    # param_path = os.path.join(os.getcwd(), 'nbmodel.txt')
    with open('nbmodel.txt', 'r') as in_:
        data = in_.readlines()
    log_priors['pos'], log_priors['neg'], log_priors['truth'], log_priors['deceptive'] = [float(x) for x in data[2].split(',')]
    for i in range(5, len(data)):
        item = data[i].split(',')
        word = item[0]
        if word not in loglikelihood:
            loglikelihood['pos'][word] = 0.0
            loglikelihood['neg'][word] = 0.0
            loglikelihood['truth'][word] = 0.0
            loglikelihood['deceptive'][word] = 0.0
        loglikelihood['pos'][word] = float(item[1])
        loglikelihood['neg'][word] = float(item[2])
        loglikelihood['truth'][word] = float(item[3])
        loglikelihood['deceptive'][word] = float(item[4].strip('\n'))
    return loglikelihood, log_priors


def classify(X, path_list):
    outfile = open('nboutput.txt', 'w')

    pos_score, neg_score, truth_score, dec_score = 0.0, 0.0, 0.0, 0.0
    truth_result = []
    pos_result = []
    for x, path in zip(X, path_list):
        counts = word_count(tokenize(x))

        for word in counts.keys():
            # print(word)
            if word in stopwords:
                continue;
            """
            log_pos = math.log10((word_counts['pos'].get(word, 0.0) + 1) / (label_count['pos'] + len(vocab)))
            log_neg = math.log10((word_counts['neg'].get(word, 0.0) + 1) / (label_count['neg'] + len(vocab)))
            log_truth = math.log10((word_counts['truth'].get(word, 0.0) + 1) / (label_count['truth'] + len(vocab)))
            log_dec = math.log10((word_counts['deceptive'].get(word, 0.0) + 1) / (label_count['deceptive'] + len(vocab)))
            # print(word)
            """
            smooth = math.log(1 / 240 + len(loglikelihood['pos']))
            pos_score += loglikelihood['pos'].get(word, smooth)
            neg_score += loglikelihood['neg'].get(word, smooth)
            truth_score += loglikelihood['truth'].get(word, smooth)
            dec_score += loglikelihood['deceptive'].get(word, smooth)

        pos_score += log_priors['pos']
        neg_score += log_priors['neg']
        truth_score += log_priors['truth']
        dec_score += log_priors['deceptive']

        if neg_score > pos_score:
            outfile.write('negative ' )
        else:
            outfile.write('positive ')

        if dec_score > truth_score:
            outfile.write('deceptive ')
        else:
            outfile.write('truthful ')

        outfile.write(path)
        outfile.write('\n')

    outfile.close()
    return pos_result, truth_result


def score():
    tp1, fp1, tn1, fn1 = 0,0,0,0
    tp2, fp2, tn2, fn2 = 0, 0,0,0
    for i in range(len(result1)):
        if result1[i] == y_test_pos[i]:
            if result1[i] == 1:
                tp1 += 1
            else:
                tn1 += 1
        else:
            if result1[i] == 1:
                fp1 += 1
            else:
                fn1 += 1
    for i in range(len(result2)):
        if result2[i] == y_test_truth[i]:
            if result2[i] == 1:
                tp2 += 1
            else:
                tn1 += 1
        else:
            if result2[i] == 1:
                fp2 += 1
            else:
                fn2 += 1
    prec1 = tp1 / (tp1 + fp1)
    rec1 = tp1 / (tp1 + fn1)
    prec2 = tp2 / (tp2 + fp2)
    rec2 = tp2 / (tp2 + fn2)
    F1_1 = 2 * ((prec1 * rec1) / (prec1 + rec1))
    F1_2 = 2 * ((prec2 * rec2) / (prec2 + rec2))
    print("-------pos/neg--------")
    print("true pos: {} | false pos: {} | false neg: {}".format(tp1, fp1, fn1))
    print("precision: {} | recall: {} | F1: {}".format(prec1, rec1, F1_1))
    print("-------truth/dec-------")
    print("precision: {} | recall: {} | F1: {}".format(prec2, rec2, F1_2))


if __name__=='__main__':
    input_path = sys.argv[-1]
    loglikelihood, log_priors = load_params()
    X_test, paths = load_data()
    stopwords = get_stopwords(X_test)
    result1, result2 = classify(X_test, paths)