Last active
June 16, 2018 17:16
-
-
Save bowbowbow/f990f3e9dc3e124e6646defab81a0442 to your computer and use it in GitHub Desktop.
TOP K=10, Mean Average Precision : 0.87821
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
import json | |
import random | |
from sklearn.feature_extraction.text import CountVectorizer | |
from sklearn.linear_model import LogisticRegression | |
from sklearn import preprocessing | |
json_data = open('./dataset.json').read() | |
tweets = json.loads(json_data) | |
tweets_train, tweets_test = [], [] | |
for tweet in tweets: | |
if random.randint(1, 10) <= 1: | |
tweets_test.append(tweet) | |
else: | |
tweets_train.append(tweet) | |
count_vec = CountVectorizer(lowercase=True, ngram_range=(1, 1)) | |
count_vec.fit([tweet['text'].replace('#', '') for tweet in tweets_train]) | |
x_train, y_train = [], [] | |
label_encoder = preprocessing.LabelEncoder() | |
for tweet in tweets_train: | |
for hashtag in tweet['hashtags']: | |
x_train.append(tweet['text']) | |
y_train.append(hashtag.lower()) | |
label_encoder.fit(y_train) | |
x_train = count_vec.transform(x_train) | |
y_train = label_encoder.transform(y_train) | |
cls = LogisticRegression() | |
cls.fit(x_train, y_train) | |
x_test = count_vec.transform([tweet['text'].replace('#', '') for tweet in tweets_test]) | |
y_test_proba = cls.predict_proba(x_test) | |
''' | |
Mean Average Precision Explanation: | |
http://sdsawtelle.github.io/blog/output/mean-average-precision-MAP-for-recommender-systems.html | |
''' | |
TOP_K = 10 | |
average_precision_sum = 0 | |
for i in range(len(y_test_proba)): | |
y_pair = [(y_test_proba[i][j], j) for j in range(len(y_test_proba[i]))] | |
ranked = list(map(lambda x: x[1], sorted(y_pair, key=lambda x: x[0], reverse=True)))[:TOP_K] | |
ranked = label_encoder.inverse_transform(ranked) | |
hashtag_map = {} | |
for hashtag in tweets_test[i]['hashtags']: | |
hashtag_map[hashtag.lower()] = True | |
hit, precision_sum = 0, 0 | |
for j in range(len(ranked)): | |
if ranked[j] in hashtag_map: | |
hit += 1 | |
precision_sum += hit / (j + 1) | |
if hit == len(tweets_test[i]['hashtags']): | |
break | |
average_precision_sum += precision_sum/len(tweets_test[i]['hashtags']) | |
print('TOP K={}, Mean Average Precision : {}'.format(TOP_K, average_precision_sum / len(tweets_test))) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment