Skip to content

Instantly share code, notes, and snippets.

@fccoelho

fccoelho/Classifica_Artigos.ipynb

Created July 17, 2014 19:14
Show Gist options
  • Save fccoelho/a62923cb69b9ac41511d to your computer and use it in GitHub Desktop.
Save fccoelho/a62923cb69b9ac41511d to your computer and use it in GitHub Desktop.
Download ZIP
experimentos de classificação de artigos do MediaCLoud
Display the source blob
Display the rendered blob
Raw
Classifica_Artigos.ipynb
{
"metadata": {
"name": "",
"signature": "sha256:c6716d2a8111d83df5a2d9ed0f53bba477bb57e99255143da939a9b2b98e9fdc"
},
"nbformat": 3,
"nbformat_minor": 0,
"worksheets": [
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#Classifica\u00e7\u00e3o de artigos do MediaCloud usando GMM e DPGMM\n",
"\n",
"Neste Notebook vamos aplicar um classificador n\u00e3o supervisionado a uma cole\u00e7\u00e3o de artigos extra\u00edda do \u00edndice do MediaCloud, usando Dirichlet Process Gaussian Mixture models, do Scikit-Learn.\n",
"http://scikit-learn.org/stable/auto_examples/mixture/plot_gmm.html#example-mixture-plot-gmm-py\n",
"\n",
"Vamos tamb\u00e9m explorar t\u00e9cnicas de Deep learning usando word2vec do pacote gensim:\n",
"http://radimrehurek.com/gensim/models/word2vec.html\n",
"\n",
"uma explica\u00e7\u00e3o mais detalhada da metodologia do word2vec pode ser encontrada aqui:\n",
"https://code.google.com/p/word2vec/\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"import pysolr\n",
"import json\n",
"import os\n",
"import gensim\n",
"import nltk # para tokeniza\u00e7\u00e3o de senten\u00e7as\n",
"from pymongo import MongoClient\n",
"from string import punctuation, digits\n",
"import bs4\n",
"from sklearn import mixture"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 72
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Primeiro vamos definir uma fun\u00e7\u00e3o para busca no \u00edndice:"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"def solr_query(index_name, query=\"\"):\n",
" \"\"\"\n",
" Perform a query in Solr server and return the documents stored in the index as JSON\n",
" \"\"\"\n",
" try:\n",
" assert index_name in ['mediacloud_articles', 'mediacloud_feeds']\n",
" except AssertionError:\n",
" return json.dumps({\"error\": \"Bad index name: {}\".format(index_name)})\n",
"\n",
" options = {\n",
" 'hl': 'true',\n",
" 'hl.fragsize': 10,\n",
" }\n",
"\n",
" server = pysolr.Solr(os.path.join(\"http://200.20.164.152:8983/solr\", index_name))\n",
" results = server.search(query, rows=1, **options)\n",
" hits = results.hits\n",
" print \"{} resultados encontrados\".format(hits)\n",
" full_results = server.search(query, rows=hits, **options)\n",
"\n",
" return json.dumps(full_results.docs)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 10
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"res = solr_query(\"mediacloud_articles\",\"copa\")\n",
"res = json.loads(res)\n",
"# retemos apenas as entradas que cont\u00e9m sum\u00e1rios\n",
"res = [d for d in res if 'summary' in d]\n",
"print \"{} resultados retidos\".format(len(res))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"24283 resultados encontrados\n",
"19820 resultados retidos"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n"
]
}
],
"prompt_number": 12
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"res[:3]"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 9,
"text": [
"[{u'_id': u'539fa8cbdcccdd3dc652dd11',\n",
" u'_version_': 1471123149631258624,\n",
" u'language': [u\"{u'code': u'pt', u'name': u'PORTUGUESE'}\"],\n",
" u'link': u'http://www.opovo.com.br/app/colunas/henriquearaujo/2014/06/12/noticiashenriquearaujo,3265569/minhas-copas.shtml',\n",
" u'links': [u\"{u'href': u'http://www.opovo.com.br/app/colunas/henriquearaujo/2014/06/12/noticiashenriquearaujo,3265569/minhas-copas.shtml', u'type': u'text/html', u'rel': u'alternate'}\"],\n",
" u'published': u'2014-06-12T04:30:00Z',\n",
" u'title': u'Minhas copas'},\n",
" {u'_id': u'539abb48dcccdd0c04f49f38',\n",
" u'_version_': 1470784514674393088,\n",
" u'language': [u\"{u'code': u'pt', u'name': u'PORTUGUESE'}\"],\n",
" u'link': u'http://www.otempo.com.br/opini%C3%A3o/tost%C3%A3o/j%C3%A1-temos-copa-1.862945',\n",
" u'links': [u\"{u'href': u'http://www.otempo.com.br/opini%C3%A3o/tost%C3%A3o/j%C3%A1-temos-copa-1.862945', u'type': u'text/html', u'rel': u'alternate'}\"],\n",
" u'published': u'2014-06-12T06:00:00Z',\n",
" u'title': u'J\\xe1 temos Copa'},\n",
" {u'_id': u'53a6689adcccdd2bfa85fee1',\n",
" u'_version_': 1471586954084614144,\n",
" u'language': [u\"{u'code': u'pt', u'name': u'PORTUGUESE'}\"],\n",
" u'link': u'http://redir.folha.com.br/redir/online/mercado/rss091/*http://www1.folha.uol.com.br/colunas/henriquemeirelles/2014/06/1474213-depois-da-copa.shtml',\n",
" u'links': [u\"{u'href': u'http://redir.folha.com.br/redir/online/mercado/rss091/*http://www1.folha.uol.com.br/colunas/henriquemeirelles/2014/06/1474213-depois-da-copa.shtml', u'type': u'text/html', u'rel': u'alternate'}\"],\n",
" u'published': u'2014-06-22T05:00:00Z',\n",
" u'summary': u'As emo\\xe7\\xf5es da Copa s\\xf3 aumentar\\xe3o nas pr\\xf3ximas semanas. Ap\\xf3s o torneio, no entanto, o foco deve se voltar \\xe0 Olimp\\xedada do Rio, e \\xe9 importante usarmos as li\\xe7\\xf5es do Mundial para uma organiza\\xe7\\xe3o vitoriosa em 2016.\\nO principal desafio da organiza\\xe7\\xe3o da Copa foi o n\\xfamero excessivo de participantes na constru\\xe7\\xe3o de est\\xe1dios, vias de transporte etc. Isso criou situa\\xe7\\xe3o complexa e, algumas vezes, confusa. Existem arenas de clubes e de governos estaduais, com fontes diversas de financiamento.\\nA mobilidade urbana envolve munic\\xedpios e Estados. As comunica\\xe7\\xf5es envolvem operadoras de telefonia e outras empresas. J\\xe1 o governo federal cuida dos aeroportos, do trafego a\\xe9reo e de outras responsabilidades.\\n<a href=\"http://redir.folha.com.br/redir/online/mercado/rss091/*http://www1.folha.uol.com.br/colunas/henriquemeirelles/2014/06/1474213-depois-da-copa.shtml\">Leia mais</a> (06/22/2014 - 02h00)',\n",
" u'title': u'Depois da Copa'}]"
]
}
],
"prompt_number": 9
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Vamos agora construir um corpus s\u00f3 com os sum\u00e1rios"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"docs = [d['summary'] for d in res]"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 16
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Antes de poder aplicar o word2vec, precisamos criar um banco de frases que s\u00e3o a mat\u00e9ria prima do word2vec. Vamos armazenar estas frases em um banco Mongodb para facilitar a itera\u00e7\u00e3o sobre as frases."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Cada Documento dividido em frases se converte na seguinte lista:"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"nltk.tokenize.sent_tokenize(docs[0])"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 22,
"text": [
"[u'As emo\\xe7\\xf5es da Copa s\\xf3 aumentar\\xe3o nas pr\\xf3ximas semanas.',\n",
" u'Ap\\xf3s o torneio, no entanto, o foco deve se voltar \\xe0 Olimp\\xedada do Rio, e \\xe9 importante usarmos as li\\xe7\\xf5es do Mundial para uma organiza\\xe7\\xe3o vitoriosa em 2016.',\n",
" u'O principal desafio da organiza\\xe7\\xe3o da Copa foi o n\\xfamero excessivo de participantes na constru\\xe7\\xe3o de est\\xe1dios, vias de transporte etc.',\n",
" u'Isso criou situa\\xe7\\xe3o complexa e, algumas vezes, confusa.',\n",
" u'Existem arenas de clubes e de governos estaduais, com fontes diversas de financiamento.',\n",
" u'A mobilidade urbana envolve munic\\xedpios e Estados.',\n",
" u'As comunica\\xe7\\xf5es envolvem operadoras de telefonia e outras empresas.',\n",
" u'J\\xe1 o governo federal cuida dos aeroportos, do trafego a\\xe9reo e de outras responsabilidades.',\n",
" u'<a href=\"http://redir.folha.com.br/redir/online/mercado/rss091/*http://www1.folha.uol.com.br/colunas/henriquemeirelles/2014/06/1474213-depois-da-copa.shtml\">Leia mais</a> (06/22/2014 - 02h00)']"
]
}
],
"prompt_number": 22
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"#criando um banco com as frases:\n",
"client = MongoClient()\n",
"db = client.word2vvec\n",
"frases = db.frases\n",
"for n,doc in enumerate(docs):\n",
" frases.insert({'doc': n, 'frases':nltk.tokenize.sent_tokenize(doc)})\n"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 21
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Agora vamos escrever um gerador que retorne uma frase de cada vez, como uma lista de tokens. mas antes vamos reduzir as palavras para min\u00fasculas."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"def get_sentences():\n",
" for doc in frases.find({}):\n",
" for f in doc['frases']:\n",
" f = bs4.BeautifulStoneSoup(f).get_text()\n",
" yield [w.strip().strip(punctuation).strip(digits).lower() for w in f.split() if w]\n",
" "
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 63
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"sentences = get_sentences()\n",
"model = gensim.models.Word2Vec(sentences, min_count=15, size=2000, workers=8)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 150
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"model[u'maracan\u00e3']"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 151,
"text": [
"array([ -2.15475768e-04, 2.41578033e-04, 2.09647973e-04, ...,\n",
" 1.02956292e-04, -9.20899838e-05, 3.91921458e-05], dtype=float32)"
]
}
],
"prompt_number": 151
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"model.most_similar(positive=[u'futebol', u'est\u00e1dio', 'copa' ], negative=['sp'], topn=10)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 152,
"text": [
"[(u'mundo', 0.07574747502803802),\n",
" (u'pol\\xedtica', 0.07265618443489075),\n",
" (u'publicado', 0.06495336443185806),\n",
" (u'galeria', 0.06415452808141708),\n",
" (u'lucro', 0.06115793064236641),\n",
" (u'feita', 0.05765655264258385),\n",
" (u'pa\\xedses', 0.053224772214889526),\n",
" (u'seu', 0.05046949163079262),\n",
" (u'dia', 0.04996431991457939),\n",
" (u'ponto', 0.048819612711668015)]"
]
}
],
"prompt_number": 152
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"model.syn0"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 153,
"text": [
"array([[ -4.14889982e-05, 1.10162247e-04, -2.49942823e-04, ...,\n",
" 1.64366429e-04, -2.34501014e-04, 2.23364128e-04],\n",
" [ 3.84889245e-05, 1.87694372e-04, 5.42827183e-05, ...,\n",
" -2.13279913e-04, 3.98685224e-05, 1.16758507e-04],\n",
" [ -7.05769853e-05, 7.68757382e-05, 3.32957025e-05, ...,\n",
" -2.36904496e-04, 2.25413998e-04, 3.49588227e-05],\n",
" ..., \n",
" [ 9.40111859e-05, -1.19246251e-04, -2.29892801e-04, ...,\n",
" -2.27589524e-04, 1.93915097e-04, -2.05169839e-04],\n",
" [ 1.17708085e-04, -9.00297382e-05, -1.95830537e-04, ...,\n",
" -1.84344288e-04, 4.59398279e-06, -5.01395880e-05],\n",
" [ -5.19479618e-05, -2.41005764e-04, 7.04437334e-05, ...,\n",
" -9.56029398e-05, -1.95253131e-04, -3.42280473e-05]], dtype=float32)"
]
}
],
"prompt_number": 153
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"scatter(model.syn0[:,1],model.syn0[:,2]);"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAAEACAYAAAB78OvLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSZIkSZIkSZIkSZIkSZIkSZIkqff9fwK0cO6Kr1tSAAAAAElFTkSuQmCC\n",
"text": [
"<matplotlib.figure.Figure at 0x7f3d2e1cea10>"
]
}
],
"prompt_number": 154
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"len(model.vocab)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 155,
"text": [
"701"
]
}
],
"prompt_number": 155
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"model.syn0.shape"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 156,
"text": [
"(701, 2000)"
]
}
],
"prompt_number": 156
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"model.vocab"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 120,
"text": [
"{u'': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6d10>,\n",
" u',': <gensim.models.word2vec.Vocab at 0x7f3d32b91990>,\n",
" u'-': <gensim.models.word2vec.Vocab at 0x7f3d2e697750>,\n",
" u'.': <gensim.models.word2vec.Vocab at 0x7f3d2efcd210>,\n",
" u'/': <gensim.models.word2vec.Vocab at 0x7f3d2f5a3a90>,\n",
" u':': <gensim.models.word2vec.Vocab at 0x7f3d315f87d0>,\n",
" u'a': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6f50>,\n",
" u'abertura': <gensim.models.word2vec.Vocab at 0x7f3d2f5a3f90>,\n",
" u'abril': <gensim.models.word2vec.Vocab at 0x7f3d2f033050>,\n",
" u'acompanhar': <gensim.models.word2vec.Vocab at 0x7f3d2d183850>,\n",
" u'acontece': <gensim.models.word2vec.Vocab at 0x7f3d2defa210>,\n",
" u'aconteceu': <gensim.models.word2vec.Vocab at 0x7f3d32b9afd0>,\n",
" u'acordo': <gensim.models.word2vec.Vocab at 0x7f3d3148cf50>,\n",
" u'adidas': <gensim.models.word2vec.Vocab at 0x7f3d31b64d90>,\n",
" u'aeroportos': <gensim.models.word2vec.Vocab at 0x7f3d305d3990>,\n",
" u'afirmou': <gensim.models.word2vec.Vocab at 0x7f3d317d6090>,\n",
" u'afp': <gensim.models.word2vec.Vocab at 0x7f3d3099e710>,\n",
" u'agora': <gensim.models.word2vec.Vocab at 0x7f3d317d60d0>,\n",
" u'ag\\xeancia': <gensim.models.word2vec.Vocab at 0x7f3d31b64290>,\n",
" u'ainda': <gensim.models.word2vec.Vocab at 0x7f3d2d452250>,\n",
" u'alegre': <gensim.models.word2vec.Vocab at 0x7f3d30208dd0>,\n",
" u'alemanha': <gensim.models.word2vec.Vocab at 0x7f3d2e6973d0>,\n",
" u'algumas': <gensim.models.word2vec.Vocab at 0x7f3d32b91f90>,\n",
" u'alguns': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6950>,\n",
" u'almo\\xe7o': <gensim.models.word2vec.Vocab at 0x7f3d32733f50>,\n",
" u'al\\xe9m': <gensim.models.word2vec.Vocab at 0x7f3d30001390>,\n",
" u'amigos': <gensim.models.word2vec.Vocab at 0x7f3d31075d50>,\n",
" u'amistoso': <gensim.models.word2vec.Vocab at 0x7f3d2f7f3c10>,\n",
" u'ano': <gensim.models.word2vec.Vocab at 0x7f3d305d3510>,\n",
" u'anos': <gensim.models.word2vec.Vocab at 0x7f3d2f5d63d0>,\n",
" u'antes': <gensim.models.word2vec.Vocab at 0x7f3d30357f50>,\n",
" u'anunciou': <gensim.models.word2vec.Vocab at 0x7f3d332f7250>,\n",
" u'ao': <gensim.models.word2vec.Vocab at 0x7f3d3074c810>,\n",
" u'aos': <gensim.models.word2vec.Vocab at 0x7f3d305d3e50>,\n",
" u'apenas': <gensim.models.word2vec.Vocab at 0x7f3d2fca4f90>,\n",
" u'apesar': <gensim.models.word2vec.Vocab at 0x7f3d306c9d50>,\n",
" u'apoio': <gensim.models.word2vec.Vocab at 0x7f3d2d6d8f50>,\n",
" u'apresentou': <gensim.models.word2vec.Vocab at 0x7f3d2e1ce250>,\n",
" u'ap\\xf3s': <gensim.models.word2vec.Vocab at 0x7f3d335df3d0>,\n",
" u'aqui': <gensim.models.word2vec.Vocab at 0x7f3d32bbc890>,\n",
" u'arena': <gensim.models.word2vec.Vocab at 0x7f3d332f7e90>,\n",
" u'argentina': <gensim.models.word2vec.Vocab at 0x7f3d2e6971d0>,\n",
" u'argentinos': <gensim.models.word2vec.Vocab at 0x7f3d3148c510>,\n",
" u'artigo': <gensim.models.word2vec.Vocab at 0x7f3d2cd6da10>,\n",
" u'as': <gensim.models.word2vec.Vocab at 0x7f3d332a9fd0>,\n",
" u'assim': <gensim.models.word2vec.Vocab at 0x7f3d2d7ff4d0>,\n",
" u'assistir': <gensim.models.word2vec.Vocab at 0x7f3d32d68790>,\n",
" u'associa\\xe7\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d32ba2550>,\n",
" u'assunto': <gensim.models.word2vec.Vocab at 0x7f3d3148c210>,\n",
" u'at': <gensim.models.word2vec.Vocab at 0x7f3d317d62d0>,\n",
" u'atacante': <gensim.models.word2vec.Vocab at 0x7f3d2fe77f50>,\n",
" u'atl\\xe9tico': <gensim.models.word2vec.Vocab at 0x7f3d315595d0>,\n",
" u'ato': <gensim.models.word2vec.Vocab at 0x7f3d31ae7e50>,\n",
" u'atrav\\xe9s': <gensim.models.word2vec.Vocab at 0x7f3d2f033dd0>,\n",
" u'atr\\xe1s': <gensim.models.word2vec.Vocab at 0x7f3d2d4522d0>,\n",
" u'atual': <gensim.models.word2vec.Vocab at 0x7f3d2db89050>,\n",
" u'atua\\xe7\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d2f477d50>,\n",
" u'at\\xe9': <gensim.models.word2vec.Vocab at 0x7f3d305e7c50>,\n",
" u'aumento': <gensim.models.word2vec.Vocab at 0x7f3d3088d950>,\n",
" u'autor': <gensim.models.word2vec.Vocab at 0x7f3d2de76490>,\n",
" u'autoridades': <gensim.models.word2vec.Vocab at 0x7f3d2e7a3190>,\n",
" u'avenida': <gensim.models.word2vec.Vocab at 0x7f3d2efbdad0>,\n",
" u'a\\xe7\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d2f04dd90>,\n",
" u'b': <gensim.models.word2vec.Vocab at 0x7f3d2e6bcd90>,\n",
" u'barcelona': <gensim.models.word2vec.Vocab at 0x7f3d2e6bc650>,\n",
" u'beira-rio': <gensim.models.word2vec.Vocab at 0x7f3d2d60ed90>,\n",
" u'belo': <gensim.models.word2vec.Vocab at 0x7f3d32da7d50>,\n",
" u'bem': <gensim.models.word2vec.Vocab at 0x7f3d31075b90>,\n",
" u'bilh\\xf5es': <gensim.models.word2vec.Vocab at 0x7f3d2f230310>,\n",
" u'blog': <gensim.models.word2vec.Vocab at 0x7f3d315c13d0>,\n",
" u'boa': <gensim.models.word2vec.Vocab at 0x7f3d305d3fd0>,\n",
" u'bola': <gensim.models.word2vec.Vocab at 0x7f3d30357fd0>,\n",
" u'brasil': <gensim.models.word2vec.Vocab at 0x7f3d317d61d0>,\n",
" u'brasil-': <gensim.models.word2vec.Vocab at 0x7f3d2feab0d0>,\n",
" u'brasileira': <gensim.models.word2vec.Vocab at 0x7f3d315c1910>,\n",
" u'brasileiras': <gensim.models.word2vec.Vocab at 0x7f3d317d6e10>,\n",
" u'brasileiro': <gensim.models.word2vec.Vocab at 0x7f3d334d8950>,\n",
" u'brasileiros': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6110>,\n",
" u'brasileir\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d310d26d0>,\n",
" u'bras\\xedlia': <gensim.models.word2vec.Vocab at 0x7f3d317d6050>,\n",
" u'brito': <gensim.models.word2vec.Vocab at 0x7f3d2cfbd490>,\n",
" u'cada': <gensim.models.word2vec.Vocab at 0x7f3d2fca47d0>,\n",
" u'camaqu\\xe3': <gensim.models.word2vec.Vocab at 0x7f3d332b7810>,\n",
" u'camar\\xf5es': <gensim.models.word2vec.Vocab at 0x7f3d2e5ba190>,\n",
" u'camisa': <gensim.models.word2vec.Vocab at 0x7f3d304412d0>,\n",
" u'campanha': <gensim.models.word2vec.Vocab at 0x7f3d332f7890>,\n",
" u'campeonato': <gensim.models.word2vec.Vocab at 0x7f3d335dfe50>,\n",
" u'campe\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d32baded0>,\n",
" u'campe\\xf5es': <gensim.models.word2vec.Vocab at 0x7f3d2debff10>,\n",
" u'campo': <gensim.models.word2vec.Vocab at 0x7f3d332f7450>,\n",
" u'campos': <gensim.models.word2vec.Vocab at 0x7f3d3074c2d0>,\n",
" u'capital': <gensim.models.word2vec.Vocab at 0x7f3d2efbdb90>,\n",
" u'carlos': <gensim.models.word2vec.Vocab at 0x7f3d32ba2ed0>,\n",
" u'carta': <gensim.models.word2vec.Vocab at 0x7f3d32761210>,\n",
" u'casa': <gensim.models.word2vec.Vocab at 0x7f3d2e668dd0>,\n",
" u'caso': <gensim.models.word2vec.Vocab at 0x7f3d2f7c2fd0>,\n",
" u'casos': <gensim.models.word2vec.Vocab at 0x7f3d33272810>,\n",
" u'castel\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d2e959610>,\n",
" u'categoria': <gensim.models.word2vec.Vocab at 0x7f3d2d452bd0>,\n",
" u'causa': <gensim.models.word2vec.Vocab at 0x7f3d2df22c90>,\n",
" u'cbf': <gensim.models.word2vec.Vocab at 0x7f3d30f61e90>,\n",
" u'cear\\xe1': <gensim.models.word2vec.Vocab at 0x7f3d2ec85dd0>,\n",
" u'central': <gensim.models.word2vec.Vocab at 0x7f3d304200d0>,\n",
" u'centro': <gensim.models.word2vec.Vocab at 0x7f3d2debf210>,\n",
" u'cerca': <gensim.models.word2vec.Vocab at 0x7f3d31fe3910>,\n",
" u'chave': <gensim.models.word2vec.Vocab at 0x7f3d30791090>,\n",
" u'chegar': <gensim.models.word2vec.Vocab at 0x7f3d2feabd50>,\n",
" u'chegou': <gensim.models.word2vec.Vocab at 0x7f3d320e1bd0>,\n",
" u'chelsea': <gensim.models.word2vec.Vocab at 0x7f3d32a3de50>,\n",
" u'chile': <gensim.models.word2vec.Vocab at 0x7f3d306c9b10>,\n",
" u'cidade': <gensim.models.word2vec.Vocab at 0x7f3d2fe17d90>,\n",
" u'cidades': <gensim.models.word2vec.Vocab at 0x7f3d305d3050>,\n",
" u'cidades-sede': <gensim.models.word2vec.Vocab at 0x7f3d316c4910>,\n",
" u'cinco': <gensim.models.word2vec.Vocab at 0x7f3d2d6d8190>,\n",
" u'civil': <gensim.models.word2vec.Vocab at 0x7f3d32741050>,\n",
" u'classifica\\xe7\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d3303e2d0>,\n",
" u'clima': <gensim.models.word2vec.Vocab at 0x7f3d31ae7210>,\n",
" u'clube': <gensim.models.word2vec.Vocab at 0x7f3d30942810>,\n",
" u'cl\\xe1ssico': <gensim.models.word2vec.Vocab at 0x7f3d300be350>,\n",
" u'cobertura': <gensim.models.word2vec.Vocab at 0x7f3d332f7090>,\n",
" u'coluna': <gensim.models.word2vec.Vocab at 0x7f3d2d452ed0>,\n",
" u'col\\xf4mbia': <gensim.models.word2vec.Vocab at 0x7f3d2e7a3210>,\n",
" u'com': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6850>,\n",
" u'comando': <gensim.models.word2vec.Vocab at 0x7f3d335dfad0>,\n",
" u'coment\\xe1rios': <gensim.models.word2vec.Vocab at 0x7f3d2da76f50>,\n",
" u'come\\xe7a': <gensim.models.word2vec.Vocab at 0x7f3d2d9ddd10>,\n",
" u'come\\xe7ou': <gensim.models.word2vec.Vocab at 0x7f3d300014d0>,\n",
" u'comiss\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d2ebc0f90>,\n",
" u'comit\\xea': <gensim.models.word2vec.Vocab at 0x7f3d2ed63e90>,\n",
" u'como': <gensim.models.word2vec.Vocab at 0x7f3d2ffebd10>,\n",
" u'competi\\xe7\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d316c4c10>,\n",
" u'completa': <gensim.models.word2vec.Vocab at 0x7f3d306c9810>,\n",
" u'confedera\\xe7\\xf5es': <gensim.models.word2vec.Vocab at 0x7f3d305d30d0>,\n",
" u'confira': <gensim.models.word2vec.Vocab at 0x7f3d3074cd50>,\n",
" u'confronto': <gensim.models.word2vec.Vocab at 0x7f3d30001190>,\n",
" u'conquista': <gensim.models.word2vec.Vocab at 0x7f3d2e6bc090>,\n",
" u'constru\\xe7\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d2efcd650>,\n",
" u'conta': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6a10>,\n",
" u'conte\\xfado': <gensim.models.word2vec.Vocab at 0x7f3d2e1ce290>,\n",
" u'continua': <gensim.models.word2vec.Vocab at 0x7f3d2e5bad10>,\n",
" u'contra': <gensim.models.word2vec.Vocab at 0x7f3d32b91ed0>,\n",
" u'convocados': <gensim.models.word2vec.Vocab at 0x7f3d30f61450>,\n",
" u'copa': <gensim.models.word2vec.Vocab at 0x7f3d317d6950>,\n",
" u'copas': <gensim.models.word2vec.Vocab at 0x7f3d335df890>,\n",
" u'corinthians': <gensim.models.word2vec.Vocab at 0x7f3d2f5a3290>,\n",
" u'costa': <gensim.models.word2vec.Vocab at 0x7f3d3303ef10>,\n",
" u'craque': <gensim.models.word2vec.Vocab at 0x7f3d32ed3c50>,\n",
" u'cristiano': <gensim.models.word2vec.Vocab at 0x7f3d2f766f50>,\n",
" u'cro\\xe1cia': <gensim.models.word2vec.Vocab at 0x7f3d30e50510>,\n",
" u'cruz': <gensim.models.word2vec.Vocab at 0x7f3d2c03b190>,\n",
" u'cruzeiro': <gensim.models.word2vec.Vocab at 0x7f3d31b64bd0>,\n",
" u'cultura': <gensim.models.word2vec.Vocab at 0x7f3d30357a50>,\n",
" u'curitiba': <gensim.models.word2vec.Vocab at 0x7f3d30f05510>,\n",
" u'c\\xe2mara': <gensim.models.word2vec.Vocab at 0x7f3d2d68b4d0>,\n",
" u'da': <gensim.models.word2vec.Vocab at 0x7f3d317d6490>,\n",
" u'dados': <gensim.models.word2vec.Vocab at 0x7f3d30441e90>,\n",
" u'dar': <gensim.models.word2vec.Vocab at 0x7f3d2e961550>,\n",
" u'das': <gensim.models.word2vec.Vocab at 0x7f3d305d3a50>,\n",
" u'data': <gensim.models.word2vec.Vocab at 0x7f3d335df690>,\n",
" u'de': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6390>,\n",
" u'decis\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d2eae2690>,\n",
" u'defesa': <gensim.models.word2vec.Vocab at 0x7f3d2ed639d0>,\n",
" u'deixar': <gensim.models.word2vec.Vocab at 0x7f3d334d8890>,\n",
" u'deixou': <gensim.models.word2vec.Vocab at 0x7f3d2f04d750>,\n",
" u'dentro': <gensim.models.word2vec.Vocab at 0x7f3d30f05390>,\n",
" u'depois': <gensim.models.word2vec.Vocab at 0x7f3d305d3e90>,\n",
" u'deputado': <gensim.models.word2vec.Vocab at 0x7f3d32b91c50>,\n",
" u'derrota': <gensim.models.word2vec.Vocab at 0x7f3d2f7665d0>,\n",
" u'derrotar': <gensim.models.word2vec.Vocab at 0x7f3d33030710>,\n",
" u'desafio': <gensim.models.word2vec.Vocab at 0x7f3d30f05250>,\n",
" u'desde': <gensim.models.word2vec.Vocab at 0x7f3d330bd3d0>,\n",
" u'desenvolvimento': <gensim.models.word2vec.Vocab at 0x7f3d2e9614d0>,\n",
" u'desporto': <gensim.models.word2vec.Vocab at 0x7f3d2de5ad90>,\n",
" u'desta': <gensim.models.word2vec.Vocab at 0x7f3d3074cc90>,\n",
" u'deste': <gensim.models.word2vec.Vocab at 0x7f3d30208050>,\n",
" u'deu': <gensim.models.word2vec.Vocab at 0x7f3d31ae7a90>,\n",
" u'deve': <gensim.models.word2vec.Vocab at 0x7f3d316c4110>,\n",
" u'devem': <gensim.models.word2vec.Vocab at 0x7f3d32b91790>,\n",
" u'devido': <gensim.models.word2vec.Vocab at 0x7f3d2d9ddc10>,\n",
" u'dezembro': <gensim.models.word2vec.Vocab at 0x7f3d2e1c8990>,\n",
" u'dia': <gensim.models.word2vec.Vocab at 0x7f3d2f5a3610>,\n",
" u'diante': <gensim.models.word2vec.Vocab at 0x7f3d30174650>,\n",
" u'dias': <gensim.models.word2vec.Vocab at 0x7f3d305d31d0>,\n",
" u'diego': <gensim.models.word2vec.Vocab at 0x7f3d2fc42990>,\n",
" u'dif\\xedcil': <gensim.models.word2vec.Vocab at 0x7f3d33030510>,\n",
" u'dilma': <gensim.models.word2vec.Vocab at 0x7f3d317d6510>,\n",
" u'direito': <gensim.models.word2vec.Vocab at 0x7f3d3239efd0>,\n",
" u'direitos': <gensim.models.word2vec.Vocab at 0x7f3d2f766f10>,\n",
" u'disputa': <gensim.models.word2vec.Vocab at 0x7f3d316c4c50>,\n",
" u'disputada': <gensim.models.word2vec.Vocab at 0x7f3d2debfe50>,\n",
" u'disse': <gensim.models.word2vec.Vocab at 0x7f3d2d6d8450>,\n",
" u'divulgada': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6b50>,\n",
" u'divulga\\xe7\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d330fced0>,\n",
" u'divulgou': <gensim.models.word2vec.Vocab at 0x7f3d2e9599d0>,\n",
" u'diz': <gensim.models.word2vec.Vocab at 0x7f3d2ec85c90>,\n",
" u'do': <gensim.models.word2vec.Vocab at 0x7f3d317d6110>,\n",
" u'dois': <gensim.models.word2vec.Vocab at 0x7f3d3148cad0>,\n",
" u'domingo': <gensim.models.word2vec.Vocab at 0x7f3d2e5ba450>,\n",
" u'dos': <gensim.models.word2vec.Vocab at 0x7f3d305d3f50>,\n",
" u'duas': <gensim.models.word2vec.Vocab at 0x7f3d30288c90>,\n",
" u'duelo': <gensim.models.word2vec.Vocab at 0x7f3d33272750>,\n",
" u'durante': <gensim.models.word2vec.Vocab at 0x7f3d33602950>,\n",
" u'e': <gensim.models.word2vec.Vocab at 0x7f3d305d3450>,\n",
" u'edi\\xe7\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d30357f10>,\n",
" u'eduardo': <gensim.models.word2vec.Vocab at 0x7f3d2eae2750>,\n",
" u'educa\\xe7\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d2ebc0290>,\n",
" u'ele': <gensim.models.word2vec.Vocab at 0x7f3d3239ef10>,\n",
" u'eleitoral': <gensim.models.word2vec.Vocab at 0x7f3d328593d0>,\n",
" u'em': <gensim.models.word2vec.Vocab at 0x7f3d2e6687d0>,\n",
" u'empate': <gensim.models.word2vec.Vocab at 0x7f3d3303e7d0>,\n",
" u'empresa': <gensim.models.word2vec.Vocab at 0x7f3d2f5a3390>,\n",
" u'empresas': <gensim.models.word2vec.Vocab at 0x7f3d2e959d50>,\n",
" u'encontro': <gensim.models.word2vec.Vocab at 0x7f3d315c1710>,\n",
" u'enfrentar': <gensim.models.word2vec.Vocab at 0x7f3d2c03bb10>,\n",
" u'enquanto': <gensim.models.word2vec.Vocab at 0x7f3d2f5d65d0>,\n",
" u'entidade': <gensim.models.word2vec.Vocab at 0x7f3d315c1610>,\n",
" u'entre': <gensim.models.word2vec.Vocab at 0x7f3d332f7610>,\n",
" u'entrevista': <gensim.models.word2vec.Vocab at 0x7f3d2f7c2c90>,\n",
" u'equipe': <gensim.models.word2vec.Vocab at 0x7f3d2febd150>,\n",
" u'equipes': <gensim.models.word2vec.Vocab at 0x7f3d316c49d0>,\n",
" u'era': <gensim.models.word2vec.Vocab at 0x7f3d2ee77510>,\n",
" u'espanha': <gensim.models.word2vec.Vocab at 0x7f3d2d7ff610>,\n",
" u'especial': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6890>,\n",
" u'espera': <gensim.models.word2vec.Vocab at 0x7f3d32761690>,\n",
" u'espn': <gensim.models.word2vec.Vocab at 0x7f3d2fca4290>,\n",
" u'esporte': <gensim.models.word2vec.Vocab at 0x7f3d30441750>,\n",
" u'esportes': <gensim.models.word2vec.Vocab at 0x7f3d330fc590>,\n",
" u'esportivo': <gensim.models.word2vec.Vocab at 0x7f3d3074c950>,\n",
" u'essa': <gensim.models.word2vec.Vocab at 0x7f3d32de9b10>,\n",
" u'esse': <gensim.models.word2vec.Vocab at 0x7f3d2df22cd0>,\n",
" u'est': <gensim.models.word2vec.Vocab at 0x7f3d317d6310>,\n",
" u'esta': <gensim.models.word2vec.Vocab at 0x7f3d32ba2210>,\n",
" u'estado': <gensim.models.word2vec.Vocab at 0x7f3d30357590>,\n",
" u'estados': <gensim.models.word2vec.Vocab at 0x7f3d2f7c2410>,\n",
" u'estadual': <gensim.models.word2vec.Vocab at 0x7f3d325f4f90>,\n",
" u'estar': <gensim.models.word2vec.Vocab at 0x7f3d2f5a3fd0>,\n",
" u'estava': <gensim.models.word2vec.Vocab at 0x7f3d30174e10>,\n",
" u'estdio': <gensim.models.word2vec.Vocab at 0x7f3d32e23fd0>,\n",
" u'estdios': <gensim.models.word2vec.Vocab at 0x7f3d32e23bd0>,\n",
" u'este': <gensim.models.word2vec.Vocab at 0x7f3d303d3890>,\n",
" u'estreia': <gensim.models.word2vec.Vocab at 0x7f3d330bde50>,\n",
" u'est\\xe1': <gensim.models.word2vec.Vocab at 0x7f3d305d3390>,\n",
" u'est\\xe1dio': <gensim.models.word2vec.Vocab at 0x7f3d3303ed90>,\n",
" u'est\\xe1dios': <gensim.models.word2vec.Vocab at 0x7f3d2fe17b10>,\n",
" u'est\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d320e16d0>,\n",
" u'eu': <gensim.models.word2vec.Vocab at 0x7f3d2ee770d0>,\n",
" u'evento': <gensim.models.word2vec.Vocab at 0x7f3d305d39d0>,\n",
" u'exemplo': <gensim.models.word2vec.Vocab at 0x7f3d3352c150>,\n",
" u'facebook': <gensim.models.word2vec.Vocab at 0x7f3d2e607110>,\n",
" u'fala': <gensim.models.word2vec.Vocab at 0x7f3d2de763d0>,\n",
" u'fase': <gensim.models.word2vec.Vocab at 0x7f3d30208450>,\n",
" u'fato': <gensim.models.word2vec.Vocab at 0x7f3d2df22b90>,\n",
" u'faz': <gensim.models.word2vec.Vocab at 0x7f3d3303e950>,\n",
" u'fazer': <gensim.models.word2vec.Vocab at 0x7f3d2efcd910>,\n",
" u'federal': <gensim.models.word2vec.Vocab at 0x7f3d2e5ba2d0>,\n",
" u'federa\\xe7\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d322b3750>,\n",
" u'feita': <gensim.models.word2vec.Vocab at 0x7f3d2dedfe90>,\n",
" u'feito': <gensim.models.word2vec.Vocab at 0x7f3d2f766450>,\n",
" u'felipe': <gensim.models.word2vec.Vocab at 0x7f3d315c1790>,\n",
" u'felip\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d2db7c1d0>,\n",
" u'fernando': <gensim.models.word2vec.Vocab at 0x7f3d2f5a3b90>,\n",
" u'festa': <gensim.models.word2vec.Vocab at 0x7f3d334d8b50>,\n",
" u'fev': <gensim.models.word2vec.Vocab at 0x7f3d2de76c10>,\n",
" u'fez': <gensim.models.word2vec.Vocab at 0x7f3d305e21d0>,\n",
" u'ficar': <gensim.models.word2vec.Vocab at 0x7f3d2fe17a10>,\n",
" u'ficou': <gensim.models.word2vec.Vocab at 0x7f3d32effc50>,\n",
" u'fifa': <gensim.models.word2vec.Vocab at 0x7f3d2fca4750>,\n",
" u'filho': <gensim.models.word2vec.Vocab at 0x7f3d3256cf10>,\n",
" u'fim': <gensim.models.word2vec.Vocab at 0x7f3d317d6890>,\n",
" u'final': <gensim.models.word2vec.Vocab at 0x7f3d2e697450>,\n",
" u'flamengo': <gensim.models.word2vec.Vocab at 0x7f3d31a90790>,\n",
" u'foi': <gensim.models.word2vec.Vocab at 0x7f3d317d6b50>,\n",
" u'folha': <gensim.models.word2vec.Vocab at 0x7f3d32ba2c50>,\n",
" u'fora': <gensim.models.word2vec.Vocab at 0x7f3d304417d0>,\n",
" u'foram': <gensim.models.word2vec.Vocab at 0x7f3d33602fd0>,\n",
" u'forma': <gensim.models.word2vec.Vocab at 0x7f3d2f5a3a50>,\n",
" u'fortaleza': <gensim.models.word2vec.Vocab at 0x7f3d332a9f50>,\n",
" u'for\\xe7a': <gensim.models.word2vec.Vocab at 0x7f3d3303eb10>,\n",
" u'for\\xe7as': <gensim.models.word2vec.Vocab at 0x7f3d2fc42b50>,\n",
" u'foto': <gensim.models.word2vec.Vocab at 0x7f3d2db89a50>,\n",
" u'fotos': <gensim.models.word2vec.Vocab at 0x7f3d32b918d0>,\n",
" u'frente': <gensim.models.word2vec.Vocab at 0x7f3d2eae2050>,\n",
" u'futebol': <gensim.models.word2vec.Vocab at 0x7f3d305d37d0>,\n",
" u'g': <gensim.models.word2vec.Vocab at 0x7f3d2f7f3310>,\n",
" u'gabinete': <gensim.models.word2vec.Vocab at 0x7f3d30357cd0>,\n",
" u'galeria': <gensim.models.word2vec.Vocab at 0x7f3d3018ead0>,\n",
" u'ganhar': <gensim.models.word2vec.Vocab at 0x7f3d30288990>,\n",
" u'ganhou': <gensim.models.word2vec.Vocab at 0x7f3d33030750>,\n",
" u'garantiu': <gensim.models.word2vec.Vocab at 0x7f3d2d7ff850>,\n",
" u'gastos': <gensim.models.word2vec.Vocab at 0x7f3d306c9dd0>,\n",
" u'gazeta': <gensim.models.word2vec.Vocab at 0x7f3d30174890>,\n",
" u'gente': <gensim.models.word2vec.Vocab at 0x7f3d2e959e10>,\n",
" u'geral': <gensim.models.word2vec.Vocab at 0x7f3d318a0390>,\n",
" u'gest\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d2d9dd310>,\n",
" u'getty': <gensim.models.word2vec.Vocab at 0x7f3d31aa6910>,\n",
" u'globo': <gensim.models.word2vec.Vocab at 0x7f3d30441b10>,\n",
" u'gol': <gensim.models.word2vec.Vocab at 0x7f3d2eae2c50>,\n",
" u'goleiro': <gensim.models.word2vec.Vocab at 0x7f3d2caa0490>,\n",
" u'gols': <gensim.models.word2vec.Vocab at 0x7f3d33602bd0>,\n",
" u'governo': <gensim.models.word2vec.Vocab at 0x7f3d305d3e10>,\n",
" u'grande': <gensim.models.word2vec.Vocab at 0x7f3d2defa5d0>,\n",
" u'grandes': <gensim.models.word2vec.Vocab at 0x7f3d305d33d0>,\n",
" u'grupo': <gensim.models.word2vec.Vocab at 0x7f3d2fca41d0>,\n",
" u'grupos': <gensim.models.word2vec.Vocab at 0x7f3d306376d0>,\n",
" u'h': <gensim.models.word2vec.Vocab at 0x7f3d2f5d64d0>,\n",
" u'henrique': <gensim.models.word2vec.Vocab at 0x7f3d317d63d0>,\n",
" u'hist\\xf3ria': <gensim.models.word2vec.Vocab at 0x7f3d2eae25d0>,\n",
" u'hoje': <gensim.models.word2vec.Vocab at 0x7f3d305d3ad0>,\n",
" u'holanda': <gensim.models.word2vec.Vocab at 0x7f3d2f830690>,\n",
" u'hora': <gensim.models.word2vec.Vocab at 0x7f3d302080d0>,\n",
" u'horas': <gensim.models.word2vec.Vocab at 0x7f3d2d452550>,\n",
" u'horizonte': <gensim.models.word2vec.Vocab at 0x7f3d2f477950>,\n",
" u'h\\xe1': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6650>,\n",
" u'ida': <gensim.models.word2vec.Vocab at 0x7f3d30420fd0>,\n",
" u'ig': <gensim.models.word2vec.Vocab at 0x7f3d32b91e10>,\n",
" u'images': <gensim.models.word2vec.Vocab at 0x7f3d31aa68d0>,\n",
" u'importante': <gensim.models.word2vec.Vocab at 0x7f3d30420090>,\n",
" u'import\\xe2ncia': <gensim.models.word2vec.Vocab at 0x7f3d2d9dd210>,\n",
" u'imprensa': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6750>,\n",
" u'informa\\xe7\\xf5es': <gensim.models.word2vec.Vocab at 0x7f3d32effc90>,\n",
" u'informou': <gensim.models.word2vec.Vocab at 0x7f3d2e697a50>,\n",
" u'inglaterra': <gensim.models.word2vec.Vocab at 0x7f3d305e7110>,\n",
" u'ingl\\xeas': <gensim.models.word2vec.Vocab at 0x7f3d3271e210>,\n",
" u'ingressos': <gensim.models.word2vec.Vocab at 0x7f3d326b0a10>,\n",
" u'instituto': <gensim.models.word2vec.Vocab at 0x7f3d2fe17e50>,\n",
" u'inter': <gensim.models.word2vec.Vocab at 0x7f3d3239e790>,\n",
" u'internacional': <gensim.models.word2vec.Vocab at 0x7f3d2ffeb750>,\n",
" u'internet': <gensim.models.word2vec.Vocab at 0x7f3d31fe3d90>,\n",
" u'in\\xedcio': <gensim.models.word2vec.Vocab at 0x7f3d305d34d0>,\n",
" u'isso': <gensim.models.word2vec.Vocab at 0x7f3d32b9a510>,\n",
" u'italiano': <gensim.models.word2vec.Vocab at 0x7f3d2f8304d0>,\n",
" u'it\\xe1lia': <gensim.models.word2vec.Vocab at 0x7f3d31c34e10>,\n",
" u'j': <gensim.models.word2vec.Vocab at 0x7f3d2fca4a90>,\n",
" u'jan': <gensim.models.word2vec.Vocab at 0x7f3d31a90d50>,\n",
" u'janeiro': <gensim.models.word2vec.Vocab at 0x7f3d2f7c2a50>,\n",
" u'jogador': <gensim.models.word2vec.Vocab at 0x7f3d3018e390>,\n",
" u'jogadores': <gensim.models.word2vec.Vocab at 0x7f3d2db7c610>,\n",
" u'jogar': <gensim.models.word2vec.Vocab at 0x7f3d306c94d0>,\n",
" u'jogo': <gensim.models.word2vec.Vocab at 0x7f3d302088d0>,\n",
" u'jogos': <gensim.models.word2vec.Vocab at 0x7f3d2ffeb050>,\n",
" u'jornal': <gensim.models.word2vec.Vocab at 0x7f3d2e256dd0>,\n",
" u'jornalista': <gensim.models.word2vec.Vocab at 0x7f3d2debf1d0>,\n",
" u'jos\\xe9': <gensim.models.word2vec.Vocab at 0x7f3d2fe17190>,\n",
" u'jo\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d2d4528d0>,\n",
" u'julho': <gensim.models.word2vec.Vocab at 0x7f3d3088d090>,\n",
" u'junho': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6090>,\n",
" u'justi\\xe7a': <gensim.models.word2vec.Vocab at 0x7f3d2efbd910>,\n",
" u'juventude': <gensim.models.word2vec.Vocab at 0x7f3d2de5a090>,\n",
" u'j\\xe1': <gensim.models.word2vec.Vocab at 0x7f3d305d3650>,\n",
" u'j\\xe9r\\xf4me': <gensim.models.word2vec.Vocab at 0x7f3d2edfe390>,\n",
" u'lado': <gensim.models.word2vec.Vocab at 0x7f3d306c9a50>,\n",
" u'lei': <gensim.models.word2vec.Vocab at 0x7f3d2f5a3150>,\n",
" u'leia': <gensim.models.word2vec.Vocab at 0x7f3d332f7c50>,\n",
" u'les\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d330bd5d0>,\n",
" u'levar': <gensim.models.word2vec.Vocab at 0x7f3d315f8ed0>,\n",
" u'libertadores': <gensim.models.word2vec.Vocab at 0x7f3d2fc42650>,\n",
" u'liga': <gensim.models.word2vec.Vocab at 0x7f3d32aa14d0>,\n",
" u'lista': <gensim.models.word2vec.Vocab at 0x7f3d30f61910>,\n",
" u'locais': <gensim.models.word2vec.Vocab at 0x7f3d305e7210>,\n",
" u'local': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6790>,\n",
" u'logo': <gensim.models.word2vec.Vocab at 0x7f3d3148c890>,\n",
" u'lucro': <gensim.models.word2vec.Vocab at 0x7f3d2f7f34d0>,\n",
" u'lugar': <gensim.models.word2vec.Vocab at 0x7f3d30791f10>,\n",
" u'luis': <gensim.models.word2vec.Vocab at 0x7f3d32ba2fd0>,\n",
" u'luiz': <gensim.models.word2vec.Vocab at 0x7f3d335df6d0>,\n",
" u'luta': <gensim.models.word2vec.Vocab at 0x7f3d2e7a30d0>,\n",
" u'madrid': <gensim.models.word2vec.Vocab at 0x7f3d3303e5d0>,\n",
" u'maio': <gensim.models.word2vec.Vocab at 0x7f3d32d63890>,\n",
" u'maior': <gensim.models.word2vec.Vocab at 0x7f3d30441810>,\n",
" u'mais': <gensim.models.word2vec.Vocab at 0x7f3d332f7410>,\n",
" u'manaus': <gensim.models.word2vec.Vocab at 0x7f3d32de9750>,\n",
" u'manh\\xe3': <gensim.models.word2vec.Vocab at 0x7f3d315c15d0>,\n",
" u'manifesta\\xe7\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d2d6d8e50>,\n",
" u'manifesta\\xe7\\xf5es': <gensim.models.word2vec.Vocab at 0x7f3d2fe77a50>,\n",
" u'mar': <gensim.models.word2vec.Vocab at 0x7f3d2debf050>,\n",
" u'maracan\\xe3': <gensim.models.word2vec.Vocab at 0x7f3d2e5ba1d0>,\n",
" u'marca': <gensim.models.word2vec.Vocab at 0x7f3d2f7c2050>,\n",
" u'marcos': <gensim.models.word2vec.Vocab at 0x7f3d2d7ff090>,\n",
" u'marcou': <gensim.models.word2vec.Vocab at 0x7f3d3256c110>,\n",
" u'mar\\xe7o': <gensim.models.word2vec.Vocab at 0x7f3d2e1ce210>,\n",
" u'mas': <gensim.models.word2vec.Vocab at 0x7f3d30441c50>,\n",
" u'meia': <gensim.models.word2vec.Vocab at 0x7f3d30942f90>,\n",
" u'meio': <gensim.models.word2vec.Vocab at 0x7f3d2fca42d0>,\n",
" u'melhor': <gensim.models.word2vec.Vocab at 0x7f3d30357110>,\n",
" u'melhores': <gensim.models.word2vec.Vocab at 0x7f3d2d7ffe90>,\n",
" u'menos': <gensim.models.word2vec.Vocab at 0x7f3d2d452890>,\n",
" u'mercado': <gensim.models.word2vec.Vocab at 0x7f3d31ae7550>,\n",
" u'meses': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6910>,\n",
" u'mesmo': <gensim.models.word2vec.Vocab at 0x7f3d304cea90>,\n",
" u'messi': <gensim.models.word2vec.Vocab at 0x7f3d3018e710>,\n",
" u'mil': <gensim.models.word2vec.Vocab at 0x7f3d332f7210>,\n",
" u'milh\\xf5es': <gensim.models.word2vec.Vocab at 0x7f3d2f7662d0>,\n",
" u'militar': <gensim.models.word2vec.Vocab at 0x7f3d3099eb50>,\n",
" u'militares': <gensim.models.word2vec.Vocab at 0x7f3d32d634d0>,\n",
" u'minha': <gensim.models.word2vec.Vocab at 0x7f3d2db89790>,\n",
" u'ministro': <gensim.models.word2vec.Vocab at 0x7f3d3088d310>,\n",
" u'minist\\xe9rio': <gensim.models.word2vec.Vocab at 0x7f3d2d9dd7d0>,\n",
" u'minutos': <gensim.models.word2vec.Vocab at 0x7f3d2dedf650>,\n",
" u'momento': <gensim.models.word2vec.Vocab at 0x7f3d2efbd790>,\n",
" u'moradores': <gensim.models.word2vec.Vocab at 0x7f3d2e959550>,\n",
" u'morte': <gensim.models.word2vec.Vocab at 0x7f3d2f04de50>,\n",
" u'mostra': <gensim.models.word2vec.Vocab at 0x7f3d3074c650>,\n",
" u'mostrou': <gensim.models.word2vec.Vocab at 0x7f3d2ee773d0>,\n",
" u'movimento': <gensim.models.word2vec.Vocab at 0x7f3d2df22fd0>,\n",
" u'movimentos': <gensim.models.word2vec.Vocab at 0x7f3d2ebc0dd0>,\n",
" u'muita': <gensim.models.word2vec.Vocab at 0x7f3d2c03bed0>,\n",
" u'muito': <gensim.models.word2vec.Vocab at 0x7f3d2fca4690>,\n",
" u'muitos': <gensim.models.word2vec.Vocab at 0x7f3d30001710>,\n",
" u'mundial': <gensim.models.word2vec.Vocab at 0x7f3d305d3810>,\n",
" u'mundo': <gensim.models.word2vec.Vocab at 0x7f3d317d6f50>,\n",
" u'municipal': <gensim.models.word2vec.Vocab at 0x7f3d32ba2050>,\n",
" u'm\\xe9xico': <gensim.models.word2vec.Vocab at 0x7f3d305d3f90>,\n",
" u'm\\xeas': <gensim.models.word2vec.Vocab at 0x7f3d320e12d0>,\n",
" u'm\\xeddia': <gensim.models.word2vec.Vocab at 0x7f3d305e7590>,\n",
" u'm\\xfasica': <gensim.models.word2vec.Vocab at 0x7f3d30208c10>,\n",
" u'na': <gensim.models.word2vec.Vocab at 0x7f3d317d6d50>,\n",
" u'nacional': <gensim.models.word2vec.Vocab at 0x7f3d2f5a3d50>,\n",
" u'nas': <gensim.models.word2vec.Vocab at 0x7f3d305d3150>,\n",
" u'natal': <gensim.models.word2vec.Vocab at 0x7f3d2fe17c10>,\n",
" u'nelson': <gensim.models.word2vec.Vocab at 0x7f3d2df41550>,\n",
" u'nem': <gensim.models.word2vec.Vocab at 0x7f3d3239e290>,\n",
" u'nessa': <gensim.models.word2vec.Vocab at 0x7f3d335884d0>,\n",
" u'nesta': <gensim.models.word2vec.Vocab at 0x7f3d332f7a90>,\n",
" u'neste': <gensim.models.word2vec.Vocab at 0x7f3d2e5bae10>,\n",
" u'neymar': <gensim.models.word2vec.Vocab at 0x7f3d32ed3250>,\n",
" u'nig\\xe9ria': <gensim.models.word2vec.Vocab at 0x7f3d3303e290>,\n",
" u'no': <gensim.models.word2vec.Vocab at 0x7f3d317d6b10>,\n",
" u'noite': <gensim.models.word2vec.Vocab at 0x7f3d2fca49d0>,\n",
" u'nordeste': <gensim.models.word2vec.Vocab at 0x7f3d332f72d0>,\n",
" u'norte': <gensim.models.word2vec.Vocab at 0x7f3d320e1090>,\n",
" u'nos': <gensim.models.word2vec.Vocab at 0x7f3d3074ce50>,\n",
" u'nossa': <gensim.models.word2vec.Vocab at 0x7f3d2c03b250>,\n",
" u'nosso': <gensim.models.word2vec.Vocab at 0x7f3d2e1ceb10>,\n",
" u'nova': <gensim.models.word2vec.Vocab at 0x7f3d32ed3cd0>,\n",
" u'novo': <gensim.models.word2vec.Vocab at 0x7f3d2df220d0>,\n",
" u'novos': <gensim.models.word2vec.Vocab at 0x7f3d2efcd850>,\n",
" u'nunca': <gensim.models.word2vec.Vocab at 0x7f3d2f246110>,\n",
" u'n\\xe1utico': <gensim.models.word2vec.Vocab at 0x7f3d2c03b2d0>,\n",
" u'n\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d2fca4390>,\n",
" u'o': <gensim.models.word2vec.Vocab at 0x7f3d317d68d0>,\n",
" u'objetivo': <gensim.models.word2vec.Vocab at 0x7f3d2e5ba910>,\n",
" u'obra': <gensim.models.word2vec.Vocab at 0x7f3d30288810>,\n",
" u'obras': <gensim.models.word2vec.Vocab at 0x7f3d305d3f10>,\n",
" u'oficial': <gensim.models.word2vec.Vocab at 0x7f3d33bc1090>,\n",
" u'oitavas': <gensim.models.word2vec.Vocab at 0x7f3d30637110>,\n",
" u'on': <gensim.models.word2vec.Vocab at 0x7f3d305e7fd0>,\n",
" u'onde': <gensim.models.word2vec.Vocab at 0x7f3d2e108ad0>,\n",
" u'online': <gensim.models.word2vec.Vocab at 0x7f3d2e9fa8d0>,\n",
" u'ontem': <gensim.models.word2vec.Vocab at 0x7f3d2fca4610>,\n",
" u'opera\\xe7\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d2fc42110>,\n",
" u'organiza\\xe7\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d317d66d0>,\n",
" u'os': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6210>,\n",
" u'ou': <gensim.models.word2vec.Vocab at 0x7f3d33bde390>,\n",
" u'outras': <gensim.models.word2vec.Vocab at 0x7f3d33c31cd0>,\n",
" u'outro': <gensim.models.word2vec.Vocab at 0x7f3d316c45d0>,\n",
" u'outros': <gensim.models.word2vec.Vocab at 0x7f3d330fc4d0>,\n",
" u'outubro': <gensim.models.word2vec.Vocab at 0x7f3d2d7ffd50>,\n",
" u'para': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6f10>,\n",
" u'parceria': <gensim.models.word2vec.Vocab at 0x7f3d2f5a3510>,\n",
" u'parte': <gensim.models.word2vec.Vocab at 0x7f3d305d3350>,\n",
" u'partida': <gensim.models.word2vec.Vocab at 0x7f3d332f7d10>,\n",
" u'partidas': <gensim.models.word2vec.Vocab at 0x7f3d32effbd0>,\n",
" u'partido': <gensim.models.word2vec.Vocab at 0x7f3d325f4d50>,\n",
" u'partir': <gensim.models.word2vec.Vocab at 0x7f3d334d8bd0>,\n",
" u'pas': <gensim.models.word2vec.Vocab at 0x7f3d2fca4ad0>,\n",
" u'passado': <gensim.models.word2vec.Vocab at 0x7f3d2f7f3710>,\n",
" u'passou': <gensim.models.word2vec.Vocab at 0x7f3d2e959f90>,\n",
" u'paulista': <gensim.models.word2vec.Vocab at 0x7f3d2efbda90>,\n",
" u'paulo': <gensim.models.word2vec.Vocab at 0x7f3d317d6b90>,\n",
" u'pa\\xeds': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6cd0>,\n",
" u'pa\\xedses': <gensim.models.word2vec.Vocab at 0x7f3d2e5bab90>,\n",
" u'pedro': <gensim.models.word2vec.Vocab at 0x7f3d30507e50>,\n",
" u'pela': <gensim.models.word2vec.Vocab at 0x7f3d332f7910>,\n",
" u'pelas': <gensim.models.word2vec.Vocab at 0x7f3d33272450>,\n",
" u'pelo': <gensim.models.word2vec.Vocab at 0x7f3d317d6e50>,\n",
" u'pelos': <gensim.models.word2vec.Vocab at 0x7f3d3256c290>,\n",
" u'pernambucano': <gensim.models.word2vec.Vocab at 0x7f3d32296ed0>,\n",
" u'pernambuco': <gensim.models.word2vec.Vocab at 0x7f3d316c4950>,\n",
" u'perto': <gensim.models.word2vec.Vocab at 0x7f3d3262a310>,\n",
" u'per\\xedodo': <gensim.models.word2vec.Vocab at 0x7f3d2f033090>,\n",
" u'pesquisa': <gensim.models.word2vec.Vocab at 0x7f3d305e7dd0>,\n",
" u'pessoas': <gensim.models.word2vec.Vocab at 0x7f3d33bc1110>,\n",
" u'pituca': <gensim.models.word2vec.Vocab at 0x7f3d328599d0>,\n",
" u'plano': <gensim.models.word2vec.Vocab at 0x7f3d2e9611d0>,\n",
" u'pode': <gensim.models.word2vec.Vocab at 0x7f3d30f053d0>,\n",
" u'podem': <gensim.models.word2vec.Vocab at 0x7f3d32aa1750>,\n",
" u'poder': <gensim.models.word2vec.Vocab at 0x7f3d3074cd10>,\n",
" u'poder\\xe1': <gensim.models.word2vec.Vocab at 0x7f3d2d7ff210>,\n",
" u'policiais': <gensim.models.word2vec.Vocab at 0x7f3d2f7c2810>,\n",
" u'pol\\xedcia': <gensim.models.word2vec.Vocab at 0x7f3d30001c10>,\n",
" u'pol\\xedtica': <gensim.models.word2vec.Vocab at 0x7f3d2e6cee10>,\n",
" u'ponto': <gensim.models.word2vec.Vocab at 0x7f3d2e1ceed0>,\n",
" u'pontos': <gensim.models.word2vec.Vocab at 0x7f3d330fccd0>,\n",
" u'popular': <gensim.models.word2vec.Vocab at 0x7f3d2ed63910>,\n",
" u'popula\\xe7\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d334d82d0>,\n",
" u'por': <gensim.models.word2vec.Vocab at 0x7f3d2fca40d0>,\n",
" u'porque': <gensim.models.word2vec.Vocab at 0x7f3d2caa0dd0>,\n",
" u'porto': <gensim.models.word2vec.Vocab at 0x7f3d30208310>,\n",
" u'portugal': <gensim.models.word2vec.Vocab at 0x7f3d2f7660d0>,\n",
" u'possibilidade': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6ad0>,\n",
" u'pouco': <gensim.models.word2vec.Vocab at 0x7f3d305d3610>,\n",
" u'povo': <gensim.models.word2vec.Vocab at 0x7f3d316c4cd0>,\n",
" u'praticamente': <gensim.models.word2vec.Vocab at 0x7f3d2f766310>,\n",
" u'prefeitura': <gensim.models.word2vec.Vocab at 0x7f3d2d9ddf50>,\n",
" u'prepara\\xe7\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d3303e510>,\n",
" u'presidenta': <gensim.models.word2vec.Vocab at 0x7f3d317d6850>,\n",
" u'presidente': <gensim.models.word2vec.Vocab at 0x7f3d305e2890>,\n",
" u'press': <gensim.models.word2vec.Vocab at 0x7f3d32a3d490>,\n",
" u'primeira': <gensim.models.word2vec.Vocab at 0x7f3d332f7990>,\n",
" u'primeiro': <gensim.models.word2vec.Vocab at 0x7f3d316c4210>,\n",
" u'primeiros': <gensim.models.word2vec.Vocab at 0x7f3d2e1cef50>,\n",
" u'principais': <gensim.models.word2vec.Vocab at 0x7f3d32b91f10>,\n",
" u'principal': <gensim.models.word2vec.Vocab at 0x7f3d2e1ce3d0>,\n",
" u'problemas': <gensim.models.word2vec.Vocab at 0x7f3d330fc1d0>,\n",
" u'processo': <gensim.models.word2vec.Vocab at 0x7f3d2e959e50>,\n",
" u'programa': <gensim.models.word2vec.Vocab at 0x7f3d2d9ddfd0>,\n",
" u'projeto': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6290>,\n",
" u'promovida': <gensim.models.word2vec.Vocab at 0x7f3d30357b50>,\n",
" u'protesto': <gensim.models.word2vec.Vocab at 0x7f3d2f7f3e10>,\n",
" u'protestos': <gensim.models.word2vec.Vocab at 0x7f3d2d7ff710>,\n",
" u'pr\\xf3xima': <gensim.models.word2vec.Vocab at 0x7f3d2fc429d0>,\n",
" u'pr\\xf3ximo': <gensim.models.word2vec.Vocab at 0x7f3d2eae29d0>,\n",
" u'pt': <gensim.models.word2vec.Vocab at 0x7f3d2f5a3450>,\n",
" u'publicado': <gensim.models.word2vec.Vocab at 0x7f3d2f5a36d0>,\n",
" u'p\\xfablica': <gensim.models.word2vec.Vocab at 0x7f3d306c9e90>,\n",
" u'p\\xfablico': <gensim.models.word2vec.Vocab at 0x7f3d30441410>,\n",
" u'p\\xfablicos': <gensim.models.word2vec.Vocab at 0x7f3d2e959650>,\n",
" u'qual': <gensim.models.word2vec.Vocab at 0x7f3d32b91610>,\n",
" u'qualquer': <gensim.models.word2vec.Vocab at 0x7f3d3239e990>,\n",
" u'quando': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6e90>,\n",
" u'quarta-feira': <gensim.models.word2vec.Vocab at 0x7f3d32e23450>,\n",
" u'quartas': <gensim.models.word2vec.Vocab at 0x7f3d304dd210>,\n",
" u'quase': <gensim.models.word2vec.Vocab at 0x7f3d305d3690>,\n",
" u'quatro': <gensim.models.word2vec.Vocab at 0x7f3d33bc11d0>,\n",
" u'que': <gensim.models.word2vec.Vocab at 0x7f3d317d6190>,\n",
" u'quem': <gensim.models.word2vec.Vocab at 0x7f3d2d183e10>,\n",
" u'quer': <gensim.models.word2vec.Vocab at 0x7f3d2e5baad0>,\n",
" u'quinta-feira': <gensim.models.word2vec.Vocab at 0x7f3d2f5a30d0>,\n",
" u'r': <gensim.models.word2vec.Vocab at 0x7f3d30420490>,\n",
" u'real': <gensim.models.word2vec.Vocab at 0x7f3d3303e650>,\n",
" u'realizada': <gensim.models.word2vec.Vocab at 0x7f3d303572d0>,\n",
" u'realizado': <gensim.models.word2vec.Vocab at 0x7f3d3074c610>,\n",
" u'realizados': <gensim.models.word2vec.Vocab at 0x7f3d2e959ad0>,\n",
" u'realiza\\xe7\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d305d3850>,\n",
" u'receber': <gensim.models.word2vec.Vocab at 0x7f3d2defa290>,\n",
" u'recebeu': <gensim.models.word2vec.Vocab at 0x7f3d330bd1d0>,\n",
" u'recife': <gensim.models.word2vec.Vocab at 0x7f3d327619d0>,\n",
" u'rede': <gensim.models.word2vec.Vocab at 0x7f3d305e7610>,\n",
" u'redes': <gensim.models.word2vec.Vocab at 0x7f3d30441490>,\n",
" u'regi\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d33bc1190>,\n",
" u'rela\\xe7\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d304dd8d0>,\n",
" u'reprodu\\xe7\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d2efbda10>,\n",
" u'resultado': <gensim.models.word2vec.Vocab at 0x7f3d3303e8d0>,\n",
" u'resultados': <gensim.models.word2vec.Vocab at 0x7f3d2fca4910>,\n",
" u'reuters': <gensim.models.word2vec.Vocab at 0x7f3d2f04d5d0>,\n",
" u'revista': <gensim.models.word2vec.Vocab at 0x7f3d2f5a3490>,\n",
" u'rica': <gensim.models.word2vec.Vocab at 0x7f3d2f32eed0>,\n",
" u'rio': <gensim.models.word2vec.Vocab at 0x7f3d303570d0>,\n",
" u'rodada': <gensim.models.word2vec.Vocab at 0x7f3d2de76210>,\n",
" u'ronaldo': <gensim.models.word2vec.Vocab at 0x7f3d2f766610>,\n",
" u'rousseff': <gensim.models.word2vec.Vocab at 0x7f3d317d6990>,\n",
" u'rs': <gensim.models.word2vec.Vocab at 0x7f3d317d6fd0>,\n",
" u'ruas': <gensim.models.word2vec.Vocab at 0x7f3d33602690>,\n",
" u's': <gensim.models.word2vec.Vocab at 0x7f3d30441fd0>,\n",
" u'salvador': <gensim.models.word2vec.Vocab at 0x7f3d2f7661d0>,\n",
" u'santa': <gensim.models.word2vec.Vocab at 0x7f3d32bad710>,\n",
" u'santos': <gensim.models.word2vec.Vocab at 0x7f3d30288650>,\n",
" u'sa\\xfade': <gensim.models.word2vec.Vocab at 0x7f3d31a902d0>,\n",
" u'scolari': <gensim.models.word2vec.Vocab at 0x7f3d2e6bcfd0>,\n",
" u'se': <gensim.models.word2vec.Vocab at 0x7f3d305d3b90>,\n",
" u'secretaria': <gensim.models.word2vec.Vocab at 0x7f3d2f5a3710>,\n",
" u'secret\\xe1rio': <gensim.models.word2vec.Vocab at 0x7f3d306466d0>,\n",
" u'secret\\xe1rio-geral': <gensim.models.word2vec.Vocab at 0x7f3d32761050>,\n",
" u'sede': <gensim.models.word2vec.Vocab at 0x7f3d315c1250>,\n",
" u'sedes': <gensim.models.word2vec.Vocab at 0x7f3d32e23890>,\n",
" u'segunda': <gensim.models.word2vec.Vocab at 0x7f3d30208f90>,\n",
" u'segunda-feira': <gensim.models.word2vec.Vocab at 0x7f3d320e1dd0>,\n",
" u'segundo': <gensim.models.word2vec.Vocab at 0x7f3d32b91710>,\n",
" u'seguran\\xe7a': <gensim.models.word2vec.Vocab at 0x7f3d302084d0>,\n",
" u'seis': <gensim.models.word2vec.Vocab at 0x7f3d2ed63c50>,\n",
" u'seja': <gensim.models.word2vec.Vocab at 0x7f3d2f7c2d90>,\n",
" u'seleo': <gensim.models.word2vec.Vocab at 0x7f3d2d7ff190>,\n",
" u'sele\\xe7\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6b10>,\n",
" u'sele\\xe7\\xf5es': <gensim.models.word2vec.Vocab at 0x7f3d316c4bd0>,\n",
" u'sem': <gensim.models.word2vec.Vocab at 0x7f3d2eae2e10>,\n",
" u'semana': <gensim.models.word2vec.Vocab at 0x7f3d2fca4f50>,\n",
" u'semifinais': <gensim.models.word2vec.Vocab at 0x7f3d2debf890>,\n",
" u'semifinal': <gensim.models.word2vec.Vocab at 0x7f3d3269b2d0>,\n",
" u'sempre': <gensim.models.word2vec.Vocab at 0x7f3d31ae73d0>,\n",
" u'sendo': <gensim.models.word2vec.Vocab at 0x7f3d2e5bac90>,\n",
" u'ser': <gensim.models.word2vec.Vocab at 0x7f3d3074c290>,\n",
" u'seria': <gensim.models.word2vec.Vocab at 0x7f3d32de9cd0>,\n",
" u'servi\\xe7os': <gensim.models.word2vec.Vocab at 0x7f3d2caa0b10>,\n",
" u'ser\\xe1': <gensim.models.word2vec.Vocab at 0x7f3d32bad350>,\n",
" u'ser\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d3264d9d0>,\n",
" u'sete': <gensim.models.word2vec.Vocab at 0x7f3d2f5d61d0>,\n",
" u'setor': <gensim.models.word2vec.Vocab at 0x7f3d2cfcd890>,\n",
" u'seu': <gensim.models.word2vec.Vocab at 0x7f3d2fca4050>,\n",
" u'seus': <gensim.models.word2vec.Vocab at 0x7f3d2fca4510>,\n",
" u'sexta-feira': <gensim.models.word2vec.Vocab at 0x7f3d32f87250>,\n",
" u'sido': <gensim.models.word2vec.Vocab at 0x7f3d2f7f3250>,\n",
" u'silva': <gensim.models.word2vec.Vocab at 0x7f3d32ba2750>,\n",
" u'site': <gensim.models.word2vec.Vocab at 0x7f3d302081d0>,\n",
" u'situa\\xe7\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d32b91dd0>,\n",
" u'so': <gensim.models.word2vec.Vocab at 0x7f3d2f5a3810>,\n",
" u'sobre': <gensim.models.word2vec.Vocab at 0x7f3d305d3a10>,\n",
" u'sociais': <gensim.models.word2vec.Vocab at 0x7f3d3074c550>,\n",
" u'social': <gensim.models.word2vec.Vocab at 0x7f3d2f5a3090>,\n",
" u'sociedade': <gensim.models.word2vec.Vocab at 0x7f3d2f830250>,\n",
" u'sorteio': <gensim.models.word2vec.Vocab at 0x7f3d2fe77310>,\n",
" u'source': <gensim.models.word2vec.Vocab at 0x7f3d32db5cd0>,\n",
" u'sp': <gensim.models.word2vec.Vocab at 0x7f3d2fe17bd0>,\n",
" u'sport': <gensim.models.word2vec.Vocab at 0x7f3d332f7850>,\n",
" u'sportv': <gensim.models.word2vec.Vocab at 0x7f3d31a90a10>,\n",
" u'sua': <gensim.models.word2vec.Vocab at 0x7f3d317d6a10>,\n",
" u'suas': <gensim.models.word2vec.Vocab at 0x7f3d30f05cd0>,\n",
" u'sucesso': <gensim.models.word2vec.Vocab at 0x7f3d317d6550>,\n",
" u'sul': <gensim.models.word2vec.Vocab at 0x7f3d30357650>,\n",
" u'su\\xe1rez': <gensim.models.word2vec.Vocab at 0x7f3d2f830950>,\n",
" u's\\xe1bado': <gensim.models.word2vec.Vocab at 0x7f3d306c97d0>,\n",
" u's\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d317d6250>,\n",
" u's\\xe9rie': <gensim.models.word2vec.Vocab at 0x7f3d330fc8d0>,\n",
" u's\\xf3': <gensim.models.word2vec.Vocab at 0x7f3d2d7ff390>,\n",
" u'tamb\\xe9m': <gensim.models.word2vec.Vocab at 0x7f3d2f5a3310>,\n",
" u'tarde': <gensim.models.word2vec.Vocab at 0x7f3d316c47d0>,\n",
" u'tem': <gensim.models.word2vec.Vocab at 0x7f3d2defa1d0>,\n",
" u'tempo': <gensim.models.word2vec.Vocab at 0x7f3d32ed3110>,\n",
" u'temporada': <gensim.models.word2vec.Vocab at 0x7f3d31ae7b90>,\n",
" u'ter': <gensim.models.word2vec.Vocab at 0x7f3d3074c0d0>,\n",
" u'tera-feira': <gensim.models.word2vec.Vocab at 0x7f3d332f7b50>,\n",
" u'terminou': <gensim.models.word2vec.Vocab at 0x7f3d3099e8d0>,\n",
" u'ter\\xe1': <gensim.models.word2vec.Vocab at 0x7f3d2efcdc50>,\n",
" u'ter\\xe7a-feira': <gensim.models.word2vec.Vocab at 0x7f3d3018ebd0>,\n",
" u'teve': <gensim.models.word2vec.Vocab at 0x7f3d306c9fd0>,\n",
" u'texto': <gensim.models.word2vec.Vocab at 0x7f3d32bbca10>,\n",
" u'the': <gensim.models.word2vec.Vocab at 0x7f3d2d6d8310>,\n",
" u'time': <gensim.models.word2vec.Vocab at 0x7f3d2e23f510>,\n",
" u'tinha': <gensim.models.word2vec.Vocab at 0x7f3d31ae7c90>,\n",
" u'todas': <gensim.models.word2vec.Vocab at 0x7f3d32d68950>,\n",
" u'todo': <gensim.models.word2vec.Vocab at 0x7f3d2fca44d0>,\n",
" u'todos': <gensim.models.word2vec.Vocab at 0x7f3d2fe17890>,\n",
" u'torcedor': <gensim.models.word2vec.Vocab at 0x7f3d32aa1f50>,\n",
" u'torcedores': <gensim.models.word2vec.Vocab at 0x7f3d32e235d0>,\n",
" u'torcida': <gensim.models.word2vec.Vocab at 0x7f3d315f8350>,\n",
" u'torneio': <gensim.models.word2vec.Vocab at 0x7f3d326b06d0>,\n",
" u'trabalhadores': <gensim.models.word2vec.Vocab at 0x7f3d2e1c80d0>,\n",
" u'trabalho': <gensim.models.word2vec.Vocab at 0x7f3d33030150>,\n",
" u'treinador': <gensim.models.word2vec.Vocab at 0x7f3d2f230250>,\n",
" u'tribunal': <gensim.models.word2vec.Vocab at 0x7f3d2d6d8810>,\n",
" u'tr\\xeas': <gensim.models.word2vec.Vocab at 0x7f3d2e1c8810>,\n",
" u'tudo': <gensim.models.word2vec.Vocab at 0x7f3d30f05dd0>,\n",
" u'turismo': <gensim.models.word2vec.Vocab at 0x7f3d2e5ba710>,\n",
" u'turistas': <gensim.models.word2vec.Vocab at 0x7f3d30357c10>,\n",
" u'tv': <gensim.models.word2vec.Vocab at 0x7f3d3074cb50>,\n",
" u'twitter': <gensim.models.word2vec.Vocab at 0x7f3d2ffeb090>,\n",
" u't\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d2fca4d10>,\n",
" u't\\xe9cnica': <gensim.models.word2vec.Vocab at 0x7f3d2f6020d0>,\n",
" u't\\xe9cnico': <gensim.models.word2vec.Vocab at 0x7f3d315c17d0>,\n",
" u't\\xeam': <gensim.models.word2vec.Vocab at 0x7f3d319980d0>,\n",
" u't\\xedtulo': <gensim.models.word2vec.Vocab at 0x7f3d2caa0710>,\n",
" u'ufc': <gensim.models.word2vec.Vocab at 0x7f3d328591d0>,\n",
" u'um': <gensim.models.word2vec.Vocab at 0x7f3d305d3c50>,\n",
" u'uma': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6e10>,\n",
" u'unidos': <gensim.models.word2vec.Vocab at 0x7f3d3088dbd0>,\n",
" u'universidade': <gensim.models.word2vec.Vocab at 0x7f3d2e5ba290>,\n",
" u'uruguai': <gensim.models.word2vec.Vocab at 0x7f3d2ed63590>,\n",
" u'vaga': <gensim.models.word2vec.Vocab at 0x7f3d30420ed0>,\n",
" u'vai': <gensim.models.word2vec.Vocab at 0x7f3d2fca4110>,\n",
" u'valcke': <gensim.models.word2vec.Vocab at 0x7f3d2f830790>,\n",
" u'valor': <gensim.models.word2vec.Vocab at 0x7f3d2f7f32d0>,\n",
" u'veja': <gensim.models.word2vec.Vocab at 0x7f3d2f5a3f50>,\n",
" u'vem': <gensim.models.word2vec.Vocab at 0x7f3d2fca43d0>,\n",
" u'vencer': <gensim.models.word2vec.Vocab at 0x7f3d2fc42710>,\n",
" u'venceu': <gensim.models.word2vec.Vocab at 0x7f3d2f766390>,\n",
" u'venda': <gensim.models.word2vec.Vocab at 0x7f3d31fe3690>,\n",
" u'ver': <gensim.models.word2vec.Vocab at 0x7f3d2f5d6490>,\n",
" u'vez': <gensim.models.word2vec.Vocab at 0x7f3d3074c690>,\n",
" u'vezes': <gensim.models.word2vec.Vocab at 0x7f3d2f766a90>,\n",
" u'vida': <gensim.models.word2vec.Vocab at 0x7f3d31b64c50>,\n",
" u'visitantes': <gensim.models.word2vec.Vocab at 0x7f3d30420890>,\n",
" u'vit\\xf3ria': <gensim.models.word2vec.Vocab at 0x7f3d2d7ff310>,\n",
" u'voc\\xea': <gensim.models.word2vec.Vocab at 0x7f3d30791610>,\n",
" u'volta': <gensim.models.word2vec.Vocab at 0x7f3d317d69d0>,\n",
" u'voos': <gensim.models.word2vec.Vocab at 0x7f3d2efcde90>,\n",
" u'v\\xe3o': <gensim.models.word2vec.Vocab at 0x7f3d316c4f90>,\n",
" u'v\\xeddeo': <gensim.models.word2vec.Vocab at 0x7f3d2f766cd0>,\n",
" u'v\\xeddeos': <gensim.models.word2vec.Vocab at 0x7f3d327332d0>,\n",
" u'x': <gensim.models.word2vec.Vocab at 0x7f3d305d3490>,\n",
" u'zona': <gensim.models.word2vec.Vocab at 0x7f3d30288190>,\n",
" u'\\xaa': <gensim.models.word2vec.Vocab at 0x7f3d30942290>,\n",
" u'\\xba': <gensim.models.word2vec.Vocab at 0x7f3d2f230d10>,\n",
" u'\\xd7': <gensim.models.word2vec.Vocab at 0x7f3d2ffe1b90>,\n",
" u'\\xe0': <gensim.models.word2vec.Vocab at 0x7f3d2e697c10>,\n",
" u'\\xe0s': <gensim.models.word2vec.Vocab at 0x7f3d2f5a3050>,\n",
" u'\\xe1frica': <gensim.models.word2vec.Vocab at 0x7f3d2f7f3b90>,\n",
" u'\\xe9': <gensim.models.word2vec.Vocab at 0x7f3d334d8d50>,\n",
" u'\\xfaltima': <gensim.models.word2vec.Vocab at 0x7f3d2fe17b90>,\n",
" u'\\xfaltimo': <gensim.models.word2vec.Vocab at 0x7f3d305e7bd0>,\n",
" u'\\xfaltimos': <gensim.models.word2vec.Vocab at 0x7f3d317d6bd0>,\n",
" u'\\u2013': <gensim.models.word2vec.Vocab at 0x7f3d317d64d0>,\n",
" u'\\u2026': <gensim.models.word2vec.Vocab at 0x7f3d304ce390>}"
]
}
],
"prompt_number": 120
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"##Criando Clusters de palavras usando o DPGMM"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"dpgmm = mixture.DPGMM(n_components=1,n_iter=30, covariance_type='full')\n",
"dpgmm.fit(model.syn0)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 157,
"text": [
"DPGMM(alpha=1.0, covariance_type='full', init_params='wmc', min_covar=None,\n",
" n_components=1, n_iter=100, params='wmc',\n",
" random_state=<mtrand.RandomState object at 0x7f3d696cfa80>, thresh=0.01,\n",
" verbose=False)"
]
}
],
"prompt_number": 157
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"dpgmm.converged_"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 147,
"text": [
"False"
]
}
],
"prompt_number": 147
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"plot(dpgmm.predict(model.syn0));"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAYYAAAEACAYAAAC3adEgAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJEmSJEmSJEmSJEmSJElS9P953X85JRl+vgAAAABJRU5ErkJggg==\n",
"text": [
"<matplotlib.figure.Figure at 0x7f3d2f2edf10>"
]
}
],
"prompt_number": 148
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"print dpgmm.means_.shape\n",
"plot(dpgmm.means_.T);"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"(1, 200)\n"
]
},
{
"metadata": {},
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAZoAAAEACAYAAACK+7BGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
"text": [
"<matplotlib.figure.Figure at 0x7f3d33543f90>"
]
}
],
"prompt_number": 149
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Agora com o GMM"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"gmm = mixture.GMM(n_components=30, covariance_type='full')\n",
"gmm.fit(model.syn0)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 137,
"text": [
"GMM(covariance_type='full', init_params='wmc', min_covar=0.001,\n",
" n_components=30, n_init=1, n_iter=100, params='wmc', random_state=None,\n",
" thresh=0.01)"
]
}
],
"prompt_number": 137
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"print gmm.means_.shape\n",
"plot(gmm.means_.T);"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"(30, 200)\n"
]
},
{
"metadata": {},
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAZoAAAEACAYAAACK+7BGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkSuQmCC\n",
"text": [
"<matplotlib.figure.Figure at 0x7f3d2f5e7550>"
]
}
],
"prompt_number": 145
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"plot(gmm.predict(model.syn0));"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAXgAAAEACAYAAAC57G0KAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSZIkSZIkSZIkSdq3/wMKnOb5agyEEwAAAABJRU5ErkJggg==\n",
"text": [
"<matplotlib.figure.Figure at 0x7f3d2ddc79d0>"
]
}
],
"prompt_number": 138
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"gmm.converged_"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 134,
"text": [
"True"
]
}
],
"prompt_number": 134
},
{
"cell_type": "code",
"collapsed": false,
"input": [],
"language": "python",
"metadata": {},
"outputs": []
}
],
"metadata": {}
}
]
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment