Skip to content

Instantly share code, notes, and snippets.

@LowriWilliams

LowriWilliams/word2vec.ipynb

Created August 16, 2020 10:53
Show Gist options
  • Save LowriWilliams/f9df0a5d745c06a7c1a540d7bb5df0d7 to your computer and use it in GitHub Desktop.
Save LowriWilliams/f9df0a5d745c06a7c1a540d7bb5df0d7 to your computer and use it in GitHub Desktop.
Download ZIP
sms_adversarial/word2vec
Display the source blob
Display the rendered blob
Raw
word2vec.ipynb
{
"cells": [
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Time taken to train word2vec model: 2.8230369091033936\n"
]
}
],
"source": [
"# Skip-gram model (sg = 1)\n",
"size = 1000\n",
"window = 3\n",
"min_count = 1\n",
"workers = 3\n",
"sg = 1\n",
"\n",
"start_time = time.time()\n",
"tokens = pd.Series(df['patterns']).values\n",
"# Train the Word2Vec Model\n",
"w2v_model = Word2Vec(tokens, min_count = min_count, size = size, workers = workers, window = window, sg = sg)\n",
"print(\"Time taken to train word2vec model: \" + str(time.time() - start_time))\n",
"\n",
"word2vec_model_file = 'word2vec_' + str(size) + '.model'\n",
"w2v_model.save(word2vec_model_file)"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Total number of words\n",
"8355\n"
]
}
],
"source": [
"# Load the model from the model file\n",
"sg_w2v_model = Word2Vec.load(word2vec_model_file)\n",
"\n",
"# Total number of the words \n",
"print(\"Total number of words\")\n",
"print(len(sg_w2v_model.wv.vocab))"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/ipykernel_launcher.py:6: DeprecationWarning: Call to deprecated `__getitem__` (Method will be removed in 4.0.0, use self.wv.__getitem__() instead).\n",
" \n",
"/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/numpy/core/fromnumeric.py:3335: RuntimeWarning: Mean of empty slice.\n",
" out=out, **kwargs)\n",
"/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/numpy/core/_methods.py:161: RuntimeWarning: invalid value encountered in double_scalars\n",
" ret = ret.dtype.type(ret / rcount)\n"
]
}
],
"source": [
"# Store the vectors for train data in following file\n",
"word2vec_filename = 'all_review_word2vec.csv'\n",
"\n",
"with open(word2vec_filename, 'w+') as word2vec_file:\n",
" for index, row in df.iterrows():\n",
" model_vector = (np.mean([sg_w2v_model[token] for token in row['patterns']], axis=0)).tolist()\n",
" \n",
" if index == 0:\n",
" header = \",\".join(str(ele) for ele in range(1000))\n",
" word2vec_file.write(header)\n",
" word2vec_file.write(\"\\n\")\n",
" # Check if the line exists else it is vector of zeros\n",
" if type(model_vector) is list: \n",
" line1 = \",\".join( [str(vector_element) for vector_element in model_vector] )\n",
" else:\n",
" line1 = \",\".join([str(0) for i in range(size)])\n",
" word2vec_file.write(line1)\n",
" word2vec_file.write('\\n')"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [
{
"data": {
"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>0</th>\n",
" <th>1</th>\n",
" <th>2</th>\n",
" <th>3</th>\n",
" <th>4</th>\n",
" <th>5</th>\n",
" <th>6</th>\n",
" <th>7</th>\n",
" <th>8</th>\n",
" <th>9</th>\n",
" <th>...</th>\n",
" <th>992</th>\n",
" <th>993</th>\n",
" <th>994</th>\n",
" <th>995</th>\n",
" <th>996</th>\n",
" <th>997</th>\n",
" <th>998</th>\n",
" <th>999</th>\n",
" <th>label_not_spam</th>\n",
" <th>label_spam</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>-0.002518</td>\n",
" <td>0.005691</td>\n",
" <td>0.028719</td>\n",
" <td>0.112553</td>\n",
" <td>0.004954</td>\n",
" <td>-0.000118</td>\n",
" <td>-0.047354</td>\n",
" <td>0.063511</td>\n",
" <td>0.015165</td>\n",
" <td>0.065314</td>\n",
" <td>...</td>\n",
" <td>0.022455</td>\n",
" <td>0.023590</td>\n",
" <td>0.049012</td>\n",
" <td>0.042182</td>\n",
" <td>0.067843</td>\n",
" <td>0.012421</td>\n",
" <td>-0.026096</td>\n",
" <td>0.048588</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>-0.003335</td>\n",
" <td>0.006316</td>\n",
" <td>0.030497</td>\n",
" <td>0.118717</td>\n",
" <td>0.004992</td>\n",
" <td>0.000358</td>\n",
" <td>-0.049837</td>\n",
" <td>0.067021</td>\n",
" <td>0.016288</td>\n",
" <td>0.069339</td>\n",
" <td>...</td>\n",
" <td>0.024083</td>\n",
" <td>0.024686</td>\n",
" <td>0.051802</td>\n",
" <td>0.044244</td>\n",
" <td>0.071909</td>\n",
" <td>0.013398</td>\n",
" <td>-0.027529</td>\n",
" <td>0.051413</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>-0.001751</td>\n",
" <td>0.004171</td>\n",
" <td>0.024952</td>\n",
" <td>0.101387</td>\n",
" <td>0.005779</td>\n",
" <td>-0.000073</td>\n",
" <td>-0.042342</td>\n",
" <td>0.057126</td>\n",
" <td>0.014099</td>\n",
" <td>0.059079</td>\n",
" <td>...</td>\n",
" <td>0.020235</td>\n",
" <td>0.020780</td>\n",
" <td>0.044079</td>\n",
" <td>0.038243</td>\n",
" <td>0.060210</td>\n",
" <td>0.011311</td>\n",
" <td>-0.023804</td>\n",
" <td>0.043236</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>-0.004231</td>\n",
" <td>0.008591</td>\n",
" <td>0.040617</td>\n",
" <td>0.158573</td>\n",
" <td>0.006669</td>\n",
" <td>0.000397</td>\n",
" <td>-0.066657</td>\n",
" <td>0.089698</td>\n",
" <td>0.021396</td>\n",
" <td>0.092416</td>\n",
" <td>...</td>\n",
" <td>0.032019</td>\n",
" <td>0.033373</td>\n",
" <td>0.069072</td>\n",
" <td>0.059283</td>\n",
" <td>0.096124</td>\n",
" <td>0.017709</td>\n",
" <td>-0.036909</td>\n",
" <td>0.068895</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>-0.003645</td>\n",
" <td>0.007694</td>\n",
" <td>0.037233</td>\n",
" <td>0.145041</td>\n",
" <td>0.005848</td>\n",
" <td>0.000056</td>\n",
" <td>-0.060643</td>\n",
" <td>0.081297</td>\n",
" <td>0.019609</td>\n",
" <td>0.084285</td>\n",
" <td>...</td>\n",
" <td>0.028998</td>\n",
" <td>0.030250</td>\n",
" <td>0.062514</td>\n",
" <td>0.054107</td>\n",
" <td>0.087722</td>\n",
" <td>0.015767</td>\n",
" <td>-0.033356</td>\n",
" <td>0.062634</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5</th>\n",
" <td>-0.002665</td>\n",
" <td>0.006103</td>\n",
" <td>0.031361</td>\n",
" <td>0.123847</td>\n",
" <td>0.005783</td>\n",
" <td>-0.000024</td>\n",
" <td>-0.051800</td>\n",
" <td>0.069786</td>\n",
" <td>0.016936</td>\n",
" <td>0.072035</td>\n",
" <td>...</td>\n",
" <td>0.024850</td>\n",
" <td>0.025738</td>\n",
" <td>0.053763</td>\n",
" <td>0.046302</td>\n",
" <td>0.074421</td>\n",
" <td>0.013556</td>\n",
" <td>-0.028769</td>\n",
" <td>0.053455</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" </tr>\n",
" <tr>\n",
" <th>6</th>\n",
" <td>-0.002668</td>\n",
" <td>0.005939</td>\n",
" <td>0.029417</td>\n",
" <td>0.115549</td>\n",
" <td>0.005098</td>\n",
" <td>-0.000145</td>\n",
" <td>-0.048460</td>\n",
" <td>0.064702</td>\n",
" <td>0.015695</td>\n",
" <td>0.067226</td>\n",
" <td>...</td>\n",
" <td>0.023183</td>\n",
" <td>0.024084</td>\n",
" <td>0.050333</td>\n",
" <td>0.043130</td>\n",
" <td>0.069733</td>\n",
" <td>0.012727</td>\n",
" <td>-0.026763</td>\n",
" <td>0.049763</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>7</th>\n",
" <td>-0.001704</td>\n",
" <td>0.003695</td>\n",
" <td>0.020013</td>\n",
" <td>0.079906</td>\n",
" <td>0.003758</td>\n",
" <td>-0.000446</td>\n",
" <td>-0.033470</td>\n",
" <td>0.044813</td>\n",
" <td>0.010834</td>\n",
" <td>0.046494</td>\n",
" <td>...</td>\n",
" <td>0.016119</td>\n",
" <td>0.016377</td>\n",
" <td>0.034478</td>\n",
" <td>0.029979</td>\n",
" <td>0.047723</td>\n",
" <td>0.008557</td>\n",
" <td>-0.018586</td>\n",
" <td>0.034297</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>8</th>\n",
" <td>-0.002183</td>\n",
" <td>0.004303</td>\n",
" <td>0.029799</td>\n",
" <td>0.123783</td>\n",
" <td>0.007773</td>\n",
" <td>-0.001676</td>\n",
" <td>-0.052463</td>\n",
" <td>0.070131</td>\n",
" <td>0.016861</td>\n",
" <td>0.072545</td>\n",
" <td>...</td>\n",
" <td>0.025195</td>\n",
" <td>0.024295</td>\n",
" <td>0.053709</td>\n",
" <td>0.046649</td>\n",
" <td>0.073422</td>\n",
" <td>0.012588</td>\n",
" <td>-0.029430</td>\n",
" <td>0.052398</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" </tr>\n",
" <tr>\n",
" <th>9</th>\n",
" <td>-0.002857</td>\n",
" <td>0.006036</td>\n",
" <td>0.040144</td>\n",
" <td>0.166509</td>\n",
" <td>0.010320</td>\n",
" <td>0.000007</td>\n",
" <td>-0.070213</td>\n",
" <td>0.093398</td>\n",
" <td>0.022464</td>\n",
" <td>0.097874</td>\n",
" <td>...</td>\n",
" <td>0.033678</td>\n",
" <td>0.034048</td>\n",
" <td>0.072506</td>\n",
" <td>0.062497</td>\n",
" <td>0.098301</td>\n",
" <td>0.018354</td>\n",
" <td>-0.039930</td>\n",
" <td>0.069753</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"<p>10 rows × 1002 columns</p>\n",
"</div>"
],
"text/plain": [
" 0 1 2 3 4 5 6 \\\n",
"0 -0.002518 0.005691 0.028719 0.112553 0.004954 -0.000118 -0.047354 \n",
"1 -0.003335 0.006316 0.030497 0.118717 0.004992 0.000358 -0.049837 \n",
"2 -0.001751 0.004171 0.024952 0.101387 0.005779 -0.000073 -0.042342 \n",
"3 -0.004231 0.008591 0.040617 0.158573 0.006669 0.000397 -0.066657 \n",
"4 -0.003645 0.007694 0.037233 0.145041 0.005848 0.000056 -0.060643 \n",
"5 -0.002665 0.006103 0.031361 0.123847 0.005783 -0.000024 -0.051800 \n",
"6 -0.002668 0.005939 0.029417 0.115549 0.005098 -0.000145 -0.048460 \n",
"7 -0.001704 0.003695 0.020013 0.079906 0.003758 -0.000446 -0.033470 \n",
"8 -0.002183 0.004303 0.029799 0.123783 0.007773 -0.001676 -0.052463 \n",
"9 -0.002857 0.006036 0.040144 0.166509 0.010320 0.000007 -0.070213 \n",
"\n",
" 7 8 9 ... 992 993 994 995 \\\n",
"0 0.063511 0.015165 0.065314 ... 0.022455 0.023590 0.049012 0.042182 \n",
"1 0.067021 0.016288 0.069339 ... 0.024083 0.024686 0.051802 0.044244 \n",
"2 0.057126 0.014099 0.059079 ... 0.020235 0.020780 0.044079 0.038243 \n",
"3 0.089698 0.021396 0.092416 ... 0.032019 0.033373 0.069072 0.059283 \n",
"4 0.081297 0.019609 0.084285 ... 0.028998 0.030250 0.062514 0.054107 \n",
"5 0.069786 0.016936 0.072035 ... 0.024850 0.025738 0.053763 0.046302 \n",
"6 0.064702 0.015695 0.067226 ... 0.023183 0.024084 0.050333 0.043130 \n",
"7 0.044813 0.010834 0.046494 ... 0.016119 0.016377 0.034478 0.029979 \n",
"8 0.070131 0.016861 0.072545 ... 0.025195 0.024295 0.053709 0.046649 \n",
"9 0.093398 0.022464 0.097874 ... 0.033678 0.034048 0.072506 0.062497 \n",
"\n",
" 996 997 998 999 label_not_spam label_spam \n",
"0 0.067843 0.012421 -0.026096 0.048588 1 0 \n",
"1 0.071909 0.013398 -0.027529 0.051413 1 0 \n",
"2 0.060210 0.011311 -0.023804 0.043236 0 1 \n",
"3 0.096124 0.017709 -0.036909 0.068895 1 0 \n",
"4 0.087722 0.015767 -0.033356 0.062634 1 0 \n",
"5 0.074421 0.013556 -0.028769 0.053455 0 1 \n",
"6 0.069733 0.012727 -0.026763 0.049763 1 0 \n",
"7 0.047723 0.008557 -0.018586 0.034297 1 0 \n",
"8 0.073422 0.012588 -0.029430 0.052398 0 1 \n",
"9 0.098301 0.018354 -0.039930 0.069753 0 1 \n",
"\n",
"[10 rows x 1002 columns]"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"word2vec_df = pd.read_csv(word2vec_filename)\n",
"word2vec_df['label'] = df['label']\n",
"\n",
"# Encode labels\n",
"word2vec_df = pd.get_dummies(word2vec_df, columns=['label'])\n",
"\n",
"display(word2vec_df.head(10))"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.2"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment