fork(162) download
copy 
  1. import pandas as pd
  2. import os
  3. from nltk.corpus import stopwords
  4. import nltk.data
  5. import logging
  6. import numpy as np  # Make sure that numpy is imported
  7. from gensim.models import Word2Vec
  8.  
  9.  
  10. from KaggleWord2VecUtility import KaggleWord2VecUtility
  11.  
  12.  
  13. if __name__ == '__main__':
  14.  
  15.     # Read data from files
  16.     train = pd.read_csv( os.path.join(os.path.dirname(__file__), 'labeledTrainData.tsv'), header=0, delimiter="\t", quoting=3)
  17.     test = pd.read_csv(os.path.join(os.path.dirname(__file__),  'testData.tsv'), header=0, delimiter="\t", quoting=3 )
  18.     unlabeled_train = pd.read_csv( os.path.join(os.path.dirname(__file__),  "unlabeledTrainData.tsv"), header=0,delimiter="\t", quoting=3 )
  19.  
  20.     # Verify the number of reviews that were read (100,000 in total)
  21.     print "Read %d labeled train reviews, %d labeled test reviews, " \
  22.      "and %d unlabeled reviews\n" % (train["review"].size,
  23.      test["review"].size, unlabeled_train["review"].size )
  24.  
  25.  
  26.  
  27.     # Load the punkt tokenizer
  28.     tokenizer = nltk.data.load('tokenizers/punkt/english.pickle')
  29.  
  30.  
  31.  
  32.     # ****** Split the labeled and unlabeled training sets into clean sentences
  33.     #
  34.     sentences = []  # Initialize an empty list of sentences
  35.  
  36.     print "Parsing sentences from training set"
  37.     for review in train["review"]:
  38.         sentences += KaggleWord2VecUtility.review_to_sentences(review, tokenizer)
  39.  
  40.     print "Parsing sentences from unlabeled set"
  41.     for review in unlabeled_train["review"]:
  42.         sentences += KaggleWord2VecUtility.review_to_sentences(review, tokenizer)
  43.  
  44.     # ****** Set parameters and train the word2vec model
  45.     #
  46.     # Import the built-in logging module and configure it so that Word2Vec
  47.     # creates nice output messages
  48.     logging.basicConfig(format='%(asctime)s : %(levelname)s : %(message)s',\
  49.         level=logging.INFO)
  50.  
  51.     # Set values for various parameters
  52.     num_features = 300    # Word vector dimensionality
  53.     min_word_count = 40   # Minimum word count
  54.     num_workers = 4       # Number of threads to run in parallel
  55.     context = 10          # Context window size
  56.     downsampling = 1e-3   # Downsample setting for frequent words
  57.  
  58.     # Initialize and train the model (this will take some time)
  59.     print "Training Word2Vec model..."
  60.     model = Word2Vec(sentences, workers=num_workers, \
  61.                 size=num_features, min_count = min_word_count, \
  62.                 window = context, sample = downsampling, seed=1)
  63.  
  64.     # If you don't plan to train the model any further, calling
  65.     # init_sims will make the model much more memory-efficient.
  66.     model.init_sims(replace=True)
  67.  
  68.     # It can be helpful to create a meaningful model name and
  69.     # save the model for later use. You can load it later using Word2Vec.load()
  70.     model_name = "300features_40minwords_10context.txt"
  71.     model.save(model_name)
  72.  
  73.     model_name = "300features_40minwords_10context.bin"
  74.     model.save(model_name)	
  75.  
  76.     print model.doesnt_match("man woman child kitchen".split())
  77.     model.doesnt_match("france england germany berlin".split())
  78.     model.doesnt_match("paris berlin london austria".split())
  79.     model.most_similar("man")
  80.     model.most_similar("queen")
  81.     model.most_similar("awful")
Runtime error #stdin #stdout #stderr 0s 7848KB
comments (?)
stdin
copy 
Standard input is empty
stdout
copy 
Standard output is empty
stderr
copy 
Traceback (most recent call last):
  File "prog.py", line 1, in <module>
ImportError: No module named pandas
https://ideone.com/9DXo4w
language:
Python (cpython 2.7.16)
created:
10 years ago
visibility:
public

Share or Embed source code

Discover > Sphere Engine API

The brand new service which powers Ideone!

Discover > IDE Widget

Widget for compiling and running the source code in a web browser!