ium_444463/evaluation.py
2022-05-06 21:05:15 +02:00

90 lines
2.4 KiB
Python

import torch
import pandas as pd
import numpy as np
from sklearn.feature_extraction.text import TfidfVectorizer
import scipy
if __name__ == "__main__":
data = pd.read_csv('fake_job_postings.csv', engine='python')
# data = data.replace(np.nan, '', regex=True)
company_profile = data["company_profile"]
company_profile = company_profile.dropna()
company_profile = np.array(company_profile)
vectorizer = TfidfVectorizer()
company_profile = vectorizer.fit_transform(company_profile)
model = torch.load('model')
data_test = pd.read_csv('data_test.csv', engine='python', header=None)
data_test = data_test.dropna()
x_test = data_test[5]
y_test = data_test[17]
x_test = np.array(x_test)
y_test = np.array(y_test)
x_test = vectorizer.transform(x_test)
x_test = torch.tensor(scipy.sparse.csr_matrix.todense(x_test)).float()
y_test = torch.tensor(y_test)
TP = []
TF = []
FP = []
FN = []
# x_test = x_test.view(x_test.size(0), -1)
model = model.eval()
print(x_test.size())
log_ps = model(x_test)
ps = torch.exp(log_ps)
top_p, top_class = ps.topk(1, dim=1)
descr = np.array(data_test[5])
for i, (x, y) in enumerate(zip(np.array(top_class), np.array(y_test.view(*top_class.shape)))):
d = descr[i]
if x == y:
if x:
TP.append(d)
else:
TF.append(d)
else:
if x:
FP.append(d)
else:
FN.append(d)
f_score = len(TP) / (len(TP) + 0.5 * (len(FP) + len(FN)))
accuracy = (len(TP) + len(TF)) / (len(TP) + len(TF) + len(FP) + len(FN))
precision = len(TP) / ( len(TP) + len(FP) )
recall = len(TP) / ( len(TP) + len(FN) )
print(f"F- score = {f_score}")
print(f"Accuracy = {accuracy}")
print(f"Precision = {precision}")
print(f"Recall = {recall}")
f = open("metrics.txt", "a")
f.write(f"F-SCORE = {f_score}\n")
f.write(f"Accuracy = {accuracy}\n")
f.write(f"Precision = {precision}\n")
f.write(f"Recall = {recall}\n")
# f.write(f"TP descriptions:")
# for i in TP:
# f.write(i+'\n')
# f.write(f"TF descriptions:")
# for i in TF:
# f.write(i+"\n")
# f.write(f"FP descriptions:")
# for i in FP:
# f.write(i+"\n")
# f.write(f"FN descriptions:")
# for i in FN:
# f.write(i+"\n")
# f.close()
a=1