-
Notifications
You must be signed in to change notification settings - Fork 0
/
readme
25 lines (18 loc) · 945 Bytes
/
readme
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
############################################################
To evaluate this project please first make sure
that you have installed
sklearn, gensim, beautifulSoup, cython
python setup.py build_ext --inplace #to build the c-extension
python docVecTrain.py #to build the doc2vec models and classifier, it will take roughly 30 mins
python app.py# to run the web app
Due to the limited timing, I didn't have a front-end.
You may enter the url into the browser
"http://[domain]:39179/submit?review="MOVIE REVIEW FROM TEST SET WOULD BE PREFERED"
a json will be returned such that 1 indicates positve feeling and 0 indicates negative feeling
The best score I have achieved in Kaggle competition
(without telling you the answer for test cases)
is 85.7% accuracy.
You can find this information at
https://www.kaggle.com/c/word2vec-nlp-tutorial/leaderboard
by searching "Weilun Du"
##############################################################