forked from zelandiya/RAKE-tutorial
-
Notifications
You must be signed in to change notification settings - Fork 0
/
optimize_rake.py
56 lines (43 loc) · 1.87 KB
/
optimize_rake.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
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
from __future__ import absolute_import
from __future__ import print_function
from six.moves import range
__author__ = 'a_medelyan'
import test_data
import rake
import sys
# reading a directory with test documents
input_dir = sys.argv[1]
# number of top ranked keywords to evaluate
top = int(sys.argv[2])
test_set = test_data.read_data(input_dir)
best_fmeasure = 0
best_vals = []
for min_char_length in range(3,8):
for max_words_length in range(3,6):
for min_keyword_frequency in range(1,7):
rake_object = rake.Rake("SmartStoplist.txt", min_char_length, max_words_length, min_keyword_frequency)
total_fmeasure = 0
for test_doc in test_set.values():
keywords = rake_object.run(test_doc.text)
num_manual_keywords = len(test_doc.keywords)
correct = 0
try:
for i in range(0,min(top, len(keywords))):
if keywords[i][0] in set(test_doc.keywords):
correct += 1
except IndexError:
print("Problem with evaluating ", keywords)
precision = correct/float(top)
recall = correct/float(num_manual_keywords)
if precision > 0 and recall > 0:
total_fmeasure += 2*precision*recall/(precision + recall)
avg_fmeasure = round(total_fmeasure*100/float(len(test_set)), 2)
if avg_fmeasure > best_fmeasure:
best_fmeasure = avg_fmeasure
best_vals = [min_char_length, max_words_length, min_keyword_frequency]
print(min_char_length, max_words_length, min_keyword_frequency, "\t", avg_fmeasure)
print("")
print("Best result at ", best_fmeasure)
print("with\tmin_char_length", best_vals[0])
print("\tmax_words_length", best_vals[1])
print("\tmin_keyword_frequency", best_vals[2])