-
-
Save jonathanlxy/840948fc5088396652bb09f35539144a to your computer and use it in GitHub Desktop.
W2V Gist for Blogpost
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
# coding: utf-8 | |
import multiprocessing | |
import pandas as pd | |
import numpy as np | |
from nltk.corpus import stopwords | |
from nltk.tokenize import RegexpTokenizer | |
from nltk.stem.porter import PorterStemmer | |
import gensim | |
def bind_desc_skill( job ): | |
if job[1] is np.nan: | |
return job[0] | |
else: | |
return ' '.join( job ) | |
def make_sentence( doc, stem = True ): | |
low_doc = doc.lower() | |
# Punctuation removal | |
tokenizer = RegexpTokenizer( r'\w+' ) | |
tokens_P = tokenizer.tokenize( low_doc ) | |
# Stopwords removal | |
custom_stopwords = stopwords.words( 'english' ) + [ 'nbsp', 'amp' ] | |
tokens_PS = [ token for token in tokens_P if token not in custom_stopwords ] | |
# Stemming | |
if stem == True: | |
stemmer = PorterStemmer() | |
sentence = [ stemmer.stem( token ) for token in tokens_PS ] | |
else: | |
sentence = tokens_PS | |
return sentence | |
jobs = pd.read_csv( 'unique_jobs.csv', encoding = 'utf-8' ) | |
jobs[ 'desc_skill' ] = jobs[ [ 'job_description', 'skills' ] ].apply( bind_desc_skill, axis = 1 ) | |
descriptions = map( lambda doc: make_sentence( doc, stem = False ), jobs[ 'desc_skill' ].tolist() ) | |
### Train Word2Vec Model | |
w2v_model = gensim.models.Word2Vec( descriptions, min_count = 2, workers = multiprocessing.cpu_count() ) | |
# Save it | |
w2v_model.save( 'my_model.word2vec' ) |
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
print '#' * 47 | |
print '#' * 5 + 'Word2Vec Skill Similarity Search Demo' + '#' * 5 | |
print '#' * 47 + '\n' | |
print 'Loading...\n' | |
import gensim | |
def custom_skill_set( skill_set ): | |
custom_skill_set = skill_set[ : ] | |
while True: | |
new_key = raw_input( 'Please add new keyword:\n' ) | |
if new_key: | |
custom_skill_set.append( new_key ) | |
else: | |
break | |
print 'skill input:' | |
print custom_skill_set | |
return custom_skill_set | |
def skill_find( custom_skill_set ): | |
print '{:^20} | {:^20}'.format( 'Word', 'Similarity' ) | |
print '-' * 45 | |
for w, f in w2v_model.most_similar( custom_skill_set, topn=20 ): | |
print '{:<20} | {:<20}'.format( w, f ) | |
w2v_model = gensim.models.Word2Vec.load( 'my_model.word2vec' ) | |
skill_set = [ 'dataiku' , 'python', 'matplotlib', 'sql', | |
'pandas' , 'numpy' , 'shiny' , 'r' , | |
'regression', 'lasso' , 'ridge' , 'pca', | |
'mongodb' , 'bash' , 'aws' , | |
'forest' , 'knn' , 'tableau' , | |
] | |
print 'Pre-set skills:\n%s\n' %skill_set | |
while True: | |
try: | |
choice = raw_input( 'Include pre-set skills? ( y | n )' ) | |
if choice == 'y': | |
custom_skill = custom_skill_set( skill_set ) | |
skill_find( custom_skill ) | |
elif choice == 'n': | |
custom_skill = custom_skill_set( [] ) | |
skill_find( custom_skill ) | |
else: | |
print 'Invalid Input' | |
print '#' * 50 + '\n' | |
except KeyError: | |
print 'This keyword did not appear in training data.' | |
print 'Please try a different word.' |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment