forgot-stopwords.ipynb

{"cells":[{"metadata":{"trusted":true},"cell_type":"code","source":"import pandas as pd\nimport os\nimport string\nimport re\nimport numpy as np\nfrom nltk.stem import WordNetLemmatizer\nimport nltk\nimport pickle\nprint('NLTK (DOWNLOAD ALL PACKAGES TO PERFORM NLP OPERATION)')\n\nprint('UNCOMMENT FOLLOWING LINE To GET NLTK DOWNLOADED')\n# nltk.download('all')\nstopword = nltk.corpus.stopwords.words('english')\nwordnet_lemmatizer = WordNetLemmatizer()\n","execution_count":6,"outputs":[{"output_type":"stream","text":"NLTK (DOWNLOAD ALL PACKAGES TO PERFORM NLP OPERATION)\nUNCOMMENT FOLLOWING LINE To GET NLTK DOWNLOADED\n","name":"stdout"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"!wget http://7c4292c7e0ca.ngrok.io/data.zip","execution_count":7,"outputs":[{"output_type":"stream","text":"--2020-08-26 08:22:30--  http://7c4292c7e0ca.ngrok.io/data.zip\nResolving 7c4292c7e0ca.ngrok.io (7c4292c7e0ca.ngrok.io)... 3.135.90.78, 2600:1f16:d83:1201::6e:1\nConnecting to 7c4292c7e0ca.ngrok.io (7c4292c7e0ca.ngrok.io)|3.135.90.78|:80... connected.\nHTTP request sent, awaiting response... 200 OK\nLength: 94843337 (90M) [application/zip]\nSaving to: ‘data.zip’\n\ndata.zip            100%[===================>]  90.45M   637KB/s    in 2m 55s  \n\n2020-08-26 08:25:27 (529 KB/s) - ‘data.zip’ saved [94843337/94843337]\n\n","name":"stdout"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"!unzip data.zip","execution_count":8,"outputs":[{"output_type":"stream","text":"Archive:  data.zip\n   creating: data/\n  inflating: data/enron_05_17_2015_with_labels_v2_100K_chunk_1_of_6.csv  \n  inflating: tfidf_content.pickle    \n  inflating: tfidf_subject.pickle    \n","name":"stdout"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"def preprocess(df):\n    \n    df = df[df.columns.drop(list(df.filter(regex='^Cat')))]\n    df = df[df['Date'] != '27/06/2001']  #removing the date\n    df[(df['Subject'] != 'RE:') & (df['Subject'] != 'FW:') & (df['Subject'] != 'Re:')]  #removing the max same subjects\n    del df['Unnamed: 0']\n    return df","execution_count":9,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"def clean_text(text):\n    text_nopunct = \"\".join([char for char in text.lower() if char not in string.punctuation])\n    tokens = re.split('\\W+', text_nopunct) #tokenize\n    words_without_stopwords = [word for word in tokens if word not in stopword] #remove stopwords from tokens\n    return [wordnet_lemmatizer.lemmatize(word, pos=\"v\") for word in words_without_stopwords]","execution_count":10,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"DATA = 'data' #https://data.world/brianray/enron-email-dataset\n\nFILENAMES = [os.path.join(DATA, filename) for filename in os.listdir(DATA)]\ndf = pd.read_csv('data/enron_05_17_2015_with_labels_v2_100K_chunk_1_of_6.csv')\ndf = preprocess(df)","execution_count":11,"outputs":[{"output_type":"stream","text":"/opt/conda/lib/python3.7/site-packages/IPython/core/interactiveshell.py:3063: DtypeWarning: Columns (9) have mixed types.Specify dtype option on import or set low_memory=False.\n  interactivity=interactivity, compiler=compiler, result=result)\n","name":"stderr"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"df = df[df['Subject'].notna()]\ndf = df[df['content'].notna()]\ndf[df['content'].isna()].shape","execution_count":12,"outputs":[{"output_type":"execute_result","execution_count":12,"data":{"text/plain":"(0, 15)"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"\ndfEmail = df[['Subject', 'content']]\ndfEmail.head()","execution_count":13,"outputs":[{"output_type":"execute_result","execution_count":13,"data":{"text/plain":"                        Subject  \\\n1                           Re:   \n2                      Re: test   \n4                     Re: Hello   \n5                     Re: Hello   \n7  Re: PRC review - phone calls   \n\n                                             content  \n1  Traveling to have a business meeting takes the...  \n2                      test successful. way to go!!!  \n4                  Let's shoot for Tuesday at 11:45.  \n5  Greg, How about either next Tuesday or Thursda...  \n7                   any morning between 10 and 11:30  ","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Subject</th>\n      <th>content</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>1</th>\n      <td>Re:</td>\n      <td>Traveling to have a business meeting takes the...</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>Re: test</td>\n      <td>test successful. way to go!!!</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>Re: Hello</td>\n      <td>Let's shoot for Tuesday at 11:45.</td>\n    </tr>\n    <tr>\n      <th>5</th>\n      <td>Re: Hello</td>\n      <td>Greg, How about either next Tuesday or Thursda...</td>\n    </tr>\n    <tr>\n      <th>7</th>\n      <td>Re: PRC review - phone calls</td>\n      <td>any morning between 10 and 11:30</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"content_vecotrizer = pickle.load(open('tfidf_content.pickle', \"rb\"))\nsubject_vecotrizer = pickle.load(open('tfidf_subject.pickle', \"rb\"))\nprint('Vectorizer Loaded!!')","execution_count":14,"outputs":[{"output_type":"stream","text":"Vectorizer Loaded!!\n","name":"stdout"},{"output_type":"stream","text":"/opt/conda/lib/python3.7/site-packages/sklearn/base.py:334: UserWarning: Trying to unpickle estimator TfidfTransformer from version 0.20.3 when using version 0.23.2. This might lead to breaking code or invalid results. Use at your own risk.\n  UserWarning)\n/opt/conda/lib/python3.7/site-packages/sklearn/base.py:334: UserWarning: Trying to unpickle estimator TfidfVectorizer from version 0.20.3 when using version 0.23.2. This might lead to breaking code or invalid results. Use at your own risk.\n  UserWarning)\n","name":"stderr"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"subject_matrix = subject_vecotrizer.transform(df['Subject'])","execution_count":15,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"dfSubject = pd.DataFrame(subject_matrix.toarray(), columns=subject_vecotrizer.get_feature_names())","execution_count":25,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"dfSubject.head()","execution_count":26,"outputs":[{"output_type":"execute_result","execution_count":26,"data":{"text/plain":"          0   00  000  00000000000000000000000000000000000weil  \\\n0  0.0  0.0  0.0  0.0                                      0.0   \n1  0.0  0.0  0.0  0.0                                      0.0   \n2  0.0  0.0  0.0  0.0                                      0.0   \n3  0.0  0.0  0.0  0.0                                      0.0   \n4  0.0  0.0  0.0  0.0                                      0.0   \n\n   000000000055936  000000000078904  0000105674  0000108727  0000110257  ...  \\\n0              0.0              0.0         0.0         0.0         0.0  ...   \n1              0.0              0.0         0.0         0.0         0.0  ...   \n2              0.0              0.0         0.0         0.0         0.0  ...   \n3              0.0              0.0         0.0         0.0         0.0  ...   \n4              0.0              0.0         0.0         0.0         0.0  ...   \n\n   zipter  zombie  zonal  zone  zooey  zoomerang  zpro  zulie  zycher  zykorie  \n0     0.0     0.0    0.0   0.0    0.0        0.0   0.0    0.0     0.0      0.0  \n1     0.0     0.0    0.0   0.0    0.0        0.0   0.0    0.0     0.0      0.0  \n2     0.0     0.0    0.0   0.0    0.0        0.0   0.0    0.0     0.0      0.0  \n3     0.0     0.0    0.0   0.0    0.0        0.0   0.0    0.0     0.0      0.0  \n4     0.0     0.0    0.0   0.0    0.0        0.0   0.0    0.0     0.0      0.0  \n\n[5 rows x 18002 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th></th>\n      <th>0</th>\n      <th>00</th>\n      <th>000</th>\n      <th>00000000000000000000000000000000000weil</th>\n      <th>000000000055936</th>\n      <th>000000000078904</th>\n      <th>0000105674</th>\n      <th>0000108727</th>\n      <th>0000110257</th>\n      <th>...</th>\n      <th>zipter</th>\n      <th>zombie</th>\n      <th>zonal</th>\n      <th>zone</th>\n      <th>zooey</th>\n      <th>zoomerang</th>\n      <th>zpro</th>\n      <th>zulie</th>\n      <th>zycher</th>\n      <th>zykorie</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>...</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>...</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>...</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>...</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>...</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n    </tr>\n  </tbody>\n</table>\n<p>5 rows × 18002 columns</p>\n</div>"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"content_matrix = content_vecotrizer.transform(df['content'])","execution_count":18,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"combine **subject** and **content** tfidf matrix"},{"metadata":{"trusted":true},"cell_type":"code","source":"type(content_matrix)","execution_count":27,"outputs":[{"output_type":"execute_result","execution_count":27,"data":{"text/plain":"scipy.sparse.csr.csr_matrix"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"import scipy.sparse as sp\n\na = sp.csr_matrix([[1,2,3],[4,5,6]])\nprint(\"a\")\nprint(a.toarray())\nprint(\"b\")\nb = sp.csr_matrix([[7,8,9],[10,11,12]])\nprint(b.toarray())\nprint(\"c\")\nc = sp.hstack((a,b))  # NOT np.vstack\nprint(c.toarray())\n","execution_count":28,"outputs":[{"output_type":"stream","text":"a\n[[1 2 3]\n [4 5 6]]\nb\n[[ 7  8  9]\n [10 11 12]]\nc\n[[ 1  2  3  7  8  9]\n [ 4  5  6 10 11 12]]\n","name":"stdout"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"matrix =  sp.hstack((subject_matrix,content_matrix))","execution_count":21,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"#Use KMeans clustering from scikit-learn\nfrom sklearn.cluster import KMeans","execution_count":22,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"#Find optimal cluster size by finding sum-of-squared-distances\n\nsosd = []\n#Run clustering for sizes 1 to 15 and capture inertia\nK = range(5,15)\nfor k in K:\n    km = KMeans(n_clusters=k, n_jobs=-1) #-1 will use all cores of CPU for computation\n    km = km.fit(matrix)\n    sosd.append(km.inertia_)\n    print(str(k) + \"processed\")\n    \nprint(\"Sum of squared distances : \" ,sosd)\n","execution_count":23,"outputs":[{"output_type":"stream","text":"/opt/conda/lib/python3.7/site-packages/sklearn/cluster/_kmeans.py:939: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n  \" removed in 0.25.\", FutureWarning)\n","name":"stderr"},{"output_type":"stream","text":"5processed\n","name":"stdout"},{"output_type":"stream","text":"/opt/conda/lib/python3.7/site-packages/sklearn/cluster/_kmeans.py:939: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n  \" removed in 0.25.\", FutureWarning)\n","name":"stderr"},{"output_type":"stream","text":"6processed\n","name":"stdout"},{"output_type":"stream","text":"/opt/conda/lib/python3.7/site-packages/sklearn/cluster/_kmeans.py:939: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n  \" removed in 0.25.\", FutureWarning)\n","name":"stderr"},{"output_type":"stream","text":"7processed\n","name":"stdout"},{"output_type":"stream","text":"/opt/conda/lib/python3.7/site-packages/sklearn/cluster/_kmeans.py:939: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n  \" removed in 0.25.\", FutureWarning)\n","name":"stderr"},{"output_type":"stream","text":"8processed\n","name":"stdout"},{"output_type":"stream","text":"/opt/conda/lib/python3.7/site-packages/sklearn/cluster/_kmeans.py:939: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n  \" removed in 0.25.\", FutureWarning)\n","name":"stderr"},{"output_type":"stream","text":"9processed\n","name":"stdout"},{"output_type":"stream","text":"/opt/conda/lib/python3.7/site-packages/sklearn/cluster/_kmeans.py:939: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n  \" removed in 0.25.\", FutureWarning)\n","name":"stderr"},{"output_type":"stream","text":"10processed\n","name":"stdout"},{"output_type":"stream","text":"/opt/conda/lib/python3.7/site-packages/sklearn/cluster/_kmeans.py:939: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n  \" removed in 0.25.\", FutureWarning)\n","name":"stderr"},{"output_type":"stream","text":"11processed\n","name":"stdout"},{"output_type":"stream","text":"/opt/conda/lib/python3.7/site-packages/sklearn/cluster/_kmeans.py:939: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n  \" removed in 0.25.\", FutureWarning)\n","name":"stderr"},{"output_type":"stream","text":"12processed\n","name":"stdout"},{"output_type":"stream","text":"/opt/conda/lib/python3.7/site-packages/sklearn/cluster/_kmeans.py:939: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n  \" removed in 0.25.\", FutureWarning)\n","name":"stderr"},{"output_type":"stream","text":"13processed\n","name":"stdout"},{"output_type":"stream","text":"/opt/conda/lib/python3.7/site-packages/sklearn/cluster/_kmeans.py:939: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n  \" removed in 0.25.\", FutureWarning)\n","name":"stderr"},{"output_type":"stream","text":"14processed\nSum of squared distances :  [181861.45844088378, 181470.28957695063, 180828.5682397169, 180484.63181013407, 180414.31224679833, 179726.5445876351, 179554.11158248264, 179015.25168092718, 179015.90429276446, 178415.35788247973]\n","name":"stdout"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"\n#Plot sosd against number of clusters\nimport matplotlib.pyplot as plt\n%matplotlib inline\nplt.plot(K, sosd, 'bx-')\nplt.xlabel('Cluster count')\nplt.ylabel('Sum_of_squared_distances')\nplt.title('Elbow Method For Optimal Cluster Size')\nplt.show()","execution_count":24,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 432x288 with 1 Axes>","image/png":"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\n"},"metadata":{"needs_background":"light"}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"#Split data into 9 clusters\nkmeans = KMeans(n_clusters=9, n_jobs=-1).fit(matrix)\n\n#get Cluster labels.\nclusters= kmeans.labels_","execution_count":30,"outputs":[{"output_type":"stream","text":"/opt/conda/lib/python3.7/site-packages/sklearn/cluster/_kmeans.py:939: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n  \" removed in 0.25.\", FutureWarning)\n","name":"stderr"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"clusters.shape","execution_count":32,"outputs":[{"output_type":"execute_result","execution_count":32,"data":{"text/plain":"(94464,)"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"len(set(clusters))","execution_count":33,"outputs":[{"output_type":"execute_result","execution_count":33,"data":{"text/plain":"9"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"pickle.dump(kmeans, open(\"kmeans.pkl\", \"wb\"))","execution_count":34,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"","execution_count":null,"outputs":[]}],"metadata":{"kernelspec":{"name":"python3","display_name":"Python 3","language":"python"},"language_info":{"name":"python","version":"3.7.6","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat":4,"nbformat_minor":4}