# check whether a file exists before read if os.path.isfile(result_item_path): with open(result_item_path, 'r') as f: # read content of query documant (doc, content) title = "query" count = 0 rst_list = [] for line in f.readlines(): if "Query" in line.split(): rst_list = [] words = line.split() title = words[2] resultTraingSetDict[title] = "" else: words = line.split() if len(words) < 1: resultTraingSetDict[title] = rst_list continue rst_list.append([words[0], words[1]]) return resultTraingSetDict results = get_result() assessment = readAssessment.get_assessment() precision_sum = 0 results = collections.OrderedDict(sorted(results.items())) for key, value in results.items(): print key precision_sum += readAssessment.precision(value , assessment[key]) print precision_sum / 16
word_probability = doc_words[
query_word] / doc_words_sum
if query_word in background_word:
background_probability = (
background_word[query_word] +
0.01) / (background_word_sum + 0.01)
else:
background_probability = (0.01) / (
background_word_sum + 0.01)
point += log(my_lambda * word_probability +
(1 - my_lambda) * background_probability)
docs_point[doc_key] = point
# sorted each doc of query by point
docs_point_list = sorted(docs_point.items(),
key=operator.itemgetter(1),
reverse=True)
AP += readAssessment.precision(docs_point_list,
assessment[q_key])
# mean average precision
mAP = AP / len(query)
print my_lambda
print mAP
lambda_test[my_lambda] = mAP
# get key with maximum value in lambda_test dictionary
max_lambda = max(lambda_test.iteritems(), key=operator.itemgetter(1))[0]
max_mAP = lambda_test[my_lambda]
# print Lambda and Max value
print "Max:"
print max_lambda, max_mAP
isBreak = raw_input("exit ?\n")
interval /= 10