MergenSoftAI/MergenSoft

1{2 "cells": [3  {4   "cell_type": "markdown",5   "metadata": {},6   "source": [7    "# LSTM"8   ]9  },10  {11   "cell_type": "code",12   "execution_count": 1,13   "metadata": {},14   "outputs": [15    {16     "name": "stdout",17     "output_type": "stream",18     "text": [19      "time: 0 ns (started: 2023-04-05 21:15:05 +03:00)\n"20     ]21    }22   ],23   "source": [24    "import numpy as np\n",25    "import pandas as pd\n",26    "import seaborn as sns\n",27    "import matplotlib.pyplot as plt\n",28    "from plotly.offline import iplot\n",29    "from keras.layers import Dropout\n",30    "from keras.models import Sequential\n",31    "from keras.callbacks import EarlyStopping\n",32    "from keras.preprocessing.text import Tokenizer\n",33    "from keras_preprocessing.sequence import pad_sequences\n",34    "from keras.layers import Dense, Embedding, LSTM, SpatialDropout1D\n",35    "from sklearn.model_selection import train_test_split\n",36    "from keras.utils.np_utils import to_categorical\n",37    "from mergen import StringAnalyzer, StringProfiling\n",38    "%load_ext autotime"39   ]40  },41  {42   "cell_type": "code",43   "execution_count": 2,44   "metadata": {},45   "outputs": [46    {47     "data": {48      "text/html": [49       "<div>\n",50       "<style scoped>\n",51       "    .dataframe tbody tr th:only-of-type {\n",52       "        vertical-align: middle;\n",53       "    }\n",54       "\n",55       "    .dataframe tbody tr th {\n",56       "        vertical-align: top;\n",57       "    }\n",58       "\n",59       "    .dataframe thead th {\n",60       "        text-align: right;\n",61       "    }\n",62       "</style>\n",63       "<table border=\"1\" class=\"dataframe\">\n",64       "  <thead>\n",65       "    <tr style=\"text-align: right;\">\n",66       "      <th></th>\n",67       "      <th>text</th>\n",68       "      <th>target</th>\n",69       "    </tr>\n",70       "  </thead>\n",71       "  <tbody>\n",72       "    <tr>\n",73       "      <th>0</th>\n",74       "      <td>çürük dişli</td>\n",75       "      <td>INSULT</td>\n",76       "    </tr>\n",77       "    <tr>\n",78       "      <th>1</th>\n",79       "      <td>adamın islama müslümanlara verdiği zararı gavu...</td>\n",80       "      <td>RACIST</td>\n",81       "    </tr>\n",82       "    <tr>\n",83       "      <th>2</th>\n",84       "      <td>erkekler zora gelmez</td>\n",85       "      <td>SEXIST</td>\n",86       "    </tr>\n",87       "    <tr>\n",88       "      <th>3</th>\n",89       "      <td>utanmazın götüne kazık sokmuşlar tıkırtı geliy...</td>\n",90       "      <td>PROFANITY</td>\n",91       "    </tr>\n",92       "    <tr>\n",93       "      <th>4</th>\n",94       "      <td>otomasyon sistemlerine doğrudan bağlanabilir</td>\n",95       "      <td>OTHER</td>\n",96       "    </tr>\n",97       "  </tbody>\n",98       "</table>\n",99       "</div>"100      ],101      "text/plain": [102       "                                                text     target\n",103       "0                                        çürük dişli     INSULT\n",104       "1  adamın islama müslümanlara verdiği zararı gavu...     RACIST\n",105       "2                               erkekler zora gelmez     SEXIST\n",106       "3  utanmazın götüne kazık sokmuşlar tıkırtı geliy...  PROFANITY\n",107       "4       otomasyon sistemlerine doğrudan bağlanabilir      OTHER"108      ]109     },110     "execution_count": 2,111     "metadata": {},112     "output_type": "execute_result"113    },114    {115     "name": "stdout",116     "output_type": "stream",117     "text": [118      "time: 1.31 s (started: 2023-04-05 21:15:05 +03:00)\n"119     ]120    }121   ],122   "source": [123    "data = pd.read_csv(\"teknofest_train_final.csv\", sep=\"|\", encoding=\"utf-8\")\n",124    "data = data[['text','target']]\n",125    "data = StringAnalyzer(data=data, columns='text').dataframe_analyze()\n",126    "data.head()"127   ]128  },129  {130   "cell_type": "code",131   "execution_count": 3,132   "metadata": {133    "pycharm": {134     "is_executing": true135    }136   },137   "outputs": [138    {139     "name": "stdout",140     "output_type": "stream",141     "text": [142      "23981\n",143      "time: 234 ms (started: 2023-04-05 21:15:07 +03:00)\n"144     ]145    }146   ],147   "source": [148    "MAX_NB_WORDS = 20000\n",149    "MAX_SEQUENCE_LENGTH = 25\n",150    "EMBEDDING_DIM = 100\n",151    "\n",152    "tokenizer = Tokenizer(num_words=MAX_NB_WORDS, filters='!\"#$%&()*+,-./:;<=>?@[\\]^_`{|}~', lower=True)\n",153    "tokenizer.fit_on_texts(data['text'].values)\n",154    "word_index = tokenizer.word_index\n",155    "print(len(word_index))"156   ]157  },158  {159   "cell_type": "code",160   "execution_count": 4,161   "metadata": {162    "pycharm": {163     "is_executing": true164    }165   },166   "outputs": [167    {168     "name": "stdout",169     "output_type": "stream",170     "text": [171      "(12617, 25)\n",172      "time: 203 ms (started: 2023-04-05 21:15:07 +03:00)\n"173     ]174    }175   ],176   "source": [177    "X = tokenizer.texts_to_sequences(data['text'].values)\n",178    "X = pad_sequences(X, maxlen=MAX_SEQUENCE_LENGTH)\n",179    "print(X.shape)"180   ]181  },182  {183   "cell_type": "code",184   "execution_count": 5,185   "metadata": {186    "pycharm": {187     "is_executing": true188    }189   },190   "outputs": [191    {192     "name": "stdout",193     "output_type": "stream",194     "text": [195      "(12617, 5)\n",196      "time: 16 ms (started: 2023-04-05 21:15:07 +03:00)\n"197     ]198    }199   ],200   "source": [201    "Y = pd.get_dummies(data['target']).values\n",202    "print(Y.shape)"203   ]204  },205  {206   "cell_type": "code",207   "execution_count": 6,208   "metadata": {209    "pycharm": {210     "is_executing": true211    }212   },213   "outputs": [214    {215     "name": "stdout",216     "output_type": "stream",217     "text": [218      "(11355, 25) (11355, 5)\n",219      "(1262, 25) (1262, 5)\n",220      "time: 0 ns (started: 2023-04-05 21:15:07 +03:00)\n"221     ]222    }223   ],224   "source": [225    "X_train, X_test, Y_train, Y_test = train_test_split(X,Y, test_size = 0.10, random_state = 42)\n",226    "print(X_train.shape,Y_train.shape)\n",227    "print(X_test.shape,Y_test.shape)"228   ]229  },230  {231   "cell_type": "code",232   "execution_count": 7,233   "metadata": {234    "pycharm": {235     "is_executing": true236    }237   },238   "outputs": [239    {240     "name": "stdout",241     "output_type": "stream",242     "text": [243      "Model: \"sequential\"\n",244      "_________________________________________________________________\n",245      " Layer (type)                Output Shape              Param #   \n",246      "=================================================================\n",247      " embedding (Embedding)       (None, 25, 100)           2000000   \n",248      "                                                                 \n",249      " spatial_dropout1d (SpatialD  (None, 25, 100)          0         \n",250      " ropout1D)                                                       \n",251      "                                                                 \n",252      " lstm (LSTM)                 (None, 100)               80400     \n",253      "                                                                 \n",254      " dense (Dense)               (None, 5)                 505       \n",255      "                                                                 \n",256      "=================================================================\n",257      "Total params: 2,080,905\n",258      "Trainable params: 2,080,905\n",259      "Non-trainable params: 0\n",260      "_________________________________________________________________\n",261      "None\n",262      "time: 328 ms (started: 2023-04-05 21:15:07 +03:00)\n"263     ]264    }265   ],266   "source": [267    "model = Sequential()\n",268    "model.add(Embedding(MAX_NB_WORDS, EMBEDDING_DIM, input_length=X.shape[1]))\n",269    "model.add(SpatialDropout1D(0.2))\n",270    "model.add(LSTM(100, dropout=0.2, recurrent_dropout=0.2))\n",271    "model.add(Dense(5, activation='softmax'))\n",272    "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",273    "print(model.summary())"274   ]275  },276  {277   "cell_type": "code",278   "execution_count": 8,279   "metadata": {280    "pycharm": {281     "is_executing": true282    }283   },284   "outputs": [285    {286     "name": "stdout",287     "output_type": "stream",288     "text": [289      "Epoch 1/5\n",290      "320/320 [==============================] - 36s 102ms/step - loss: 1.1623 - accuracy: 0.5266 - val_loss: 0.7300 - val_accuracy: 0.7412\n",291      "Epoch 2/5\n",292      "320/320 [==============================] - 29s 91ms/step - loss: 0.3905 - accuracy: 0.8709 - val_loss: 0.5604 - val_accuracy: 0.7958\n",293      "Epoch 3/5\n",294      "320/320 [==============================] - 29s 90ms/step - loss: 0.1497 - accuracy: 0.9527 - val_loss: 0.6357 - val_accuracy: 0.7940\n",295      "Epoch 4/5\n",296      "320/320 [==============================] - 29s 90ms/step - loss: 0.0862 - accuracy: 0.9699 - val_loss: 0.6369 - val_accuracy: 0.7887\n",297      "Epoch 5/5\n",298      "320/320 [==============================] - 29s 90ms/step - loss: 0.0668 - accuracy: 0.9767 - val_loss: 0.7469 - val_accuracy: 0.7861\n",299      "time: 2min 31s (started: 2023-04-05 21:15:08 +03:00)\n"300     ]301    }302   ],303   "source": [304    "epochs = 5\n",305    "batch_size = 32\n",306    "\n",307    "history = model.fit(X_train, Y_train, epochs=epochs, batch_size=batch_size,validation_split=0.1,callbacks=[EarlyStopping(monitor='val_loss', patience=3, min_delta=0.0001)])"308   ]309  },310  {311   "cell_type": "code",312   "execution_count": 9,313   "metadata": {314    "pycharm": {315     "is_executing": true316    }317   },318   "outputs": [319    {320     "name": "stdout",321     "output_type": "stream",322     "text": [323      "40/40 [==============================] - 1s 14ms/step - loss: 0.7927 - accuracy: 0.7876\n",324      "Test set\n",325      "  Loss: 0.793\n",326      "  Accuracy: 0.788\n",327      "time: 656 ms (started: 2023-04-05 21:17:39 +03:00)\n"328     ]329    }330   ],331   "source": [332    "accr = model.evaluate(X_test,Y_test)\n",333    "print('Test set\\n  Loss: {:0.3f}\\n  Accuracy: {:0.3f}'.format(accr[0],accr[1]))"334   ]335  },336  {337   "cell_type": "code",338   "execution_count": 11,339   "metadata": {340    "pycharm": {341     "is_executing": true342    }343   },344   "outputs": [345    {346     "name": "stdout",347     "output_type": "stream",348     "text": [349      "355/355 [==============================] - 3s 8ms/step\n",350      "40/40 [==============================] - 0s 9ms/step\n",351      "Train F1 score:  0.9628523307504523\n",352      "Test F1 score:  0.7921330050595278\n",353      "time: 3.55 s (started: 2023-04-05 21:18:35 +03:00)\n"354     ]355    }356   ],357   "source": [358    "from sklearn.metrics import f1_score\n",359    "\n",360    "# Eğitim verilerinin tahmini\n",361    "y_pred_train = model.predict(X_train)\n",362    "y_pred_train = np.argmax(y_pred_train, axis=1)\n",363    "\n",364    "# F1 skoru hesaplama\n",365    "f1_train = f1_score(np.argmax(Y_train, axis=1), y_pred_train, average='weighted')\n",366    "\n",367    "# Test verilerinin tahmini\n",368    "y_pred_test = model.predict(X_test)\n",369    "y_pred_test = np.argmax(y_pred_test, axis=1)\n",370    "\n",371    "# F1 skoru hesaplama\n",372    "f1_test = f1_score(np.argmax(Y_test, axis=1), y_pred_test, average='weighted')\n",373    "\n",374    "print(\"Train F1 score: \", f1_train)\n",375    "print(\"Test F1 score: \", f1_test)"376   ]377  },378  {379   "cell_type": "code",380   "execution_count": 12,381   "metadata": {382    "pycharm": {383     "is_executing": true384    }385   },386   "outputs": [387    {388     "name": "stdout",389     "output_type": "stream",390     "text": [391      "40/40 [==============================] - 0s 9ms/step\n"392     ]393    },394    {395     "data": {396      "image/png": "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\n",397      "text/plain": [398       "<Figure size 640x480 with 2 Axes>"399      ]400     },401     "metadata": {},402     "output_type": "display_data"403    },404    {405     "name": "stdout",406     "output_type": "stream",407     "text": [408      "time: 735 ms (started: 2023-04-05 21:18:41 +03:00)\n"409     ]410    }411   ],412   "source": [413    "from sklearn.metrics import confusion_matrix\n",414    "import seaborn as sns\n",415    "import matplotlib.pyplot as plt\n",416    "# kategori isimleri\n",417    "categories = data['target'].unique().tolist()\n",418    "# modelin tahminleri\n",419    "y_pred = model.predict(X_test)\n",420    "y_pred = np.argmax(y_pred, axis=1)\n",421    "# gerçek sınıflar\n",422    "y_true = np.argmax(Y_test, axis=1)\n",423    "# confusion matrix\n",424    "cm = confusion_matrix(y_true, y_pred)\n",425    "# confusion matrix grafiği\n",426    "sns.heatmap(cm, annot=True, cmap='Blues', fmt='g', xticklabels=categories, yticklabels=categories)\n",427    "plt.xlabel('Predicted')\n",428    "plt.ylabel('True')\n",429    "plt.show()"430   ]431  },432  {433   "cell_type": "code",434   "execution_count": 13,435   "metadata": {436    "pycharm": {437     "is_executing": true438    }439   },440   "outputs": [441    {442     "name": "stdout",443     "output_type": "stream",444     "text": [445      "40/40 [==============================] - 0s 9ms/step\n"446     ]447    },448    {449     "data": {450      "image/png": "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\n",451      "text/plain": [452       "<Figure size 1000x600 with 1 Axes>"453      ]454     },455     "metadata": {},456     "output_type": "display_data"457    },458    {459     "name": "stdout",460     "output_type": "stream",461     "text": [462      "time: 593 ms (started: 2023-04-05 21:18:44 +03:00)\n"463     ]464    }465   ],466   "source": [467    "from sklearn.metrics import classification_report\n",468    "\n",469    "# Sınıflandırma etiketlerini tahmin et\n",470    "Y_pred = model.predict(X_test)\n",471    "\n",472    "# En yüksek olasılığa sahip sınıf etiketlerini seç\n",473    "Y_pred_classes = np.argmax(Y_pred, axis=1)\n",474    "Y_true = np.argmax(Y_test, axis=1)\n",475    "\n",476    "# Sınıf bazında doğruluk, hassasiyet ve F1 skorlarını hesapla\n",477    "report = classification_report(Y_true, Y_pred_classes, target_names=['INSULT', 'RACIST', 'SEXIST', 'PROFANITY', 'OTHER'], output_dict=True)\n",478    "\n",479    "# Bar grafiği oluştur\n",480    "fig, ax = plt.subplots(figsize=(10, 6))\n",481    "class_names = list(report.keys())[:-3] # Son üç satır hassasiyet, doğruluk ve F1 Skoru toplamları olduğundan bunları çıkardık\n",482    "metrics = ['precision', 'recall', 'f1-score']\n",483    "for i, metric in enumerate(metrics):\n",484    "    ax.bar([j+(i*0.8/3) for j in range(1, len(class_names)+1)], [report[class_name][metric] for class_name in class_names], width=0.8/3, label=metric)\n",485    "ax.set_xticks(range(1, len(class_names)+1))\n",486    "ax.set_xticklabels(class_names)\n",487    "ax.set_ylim([0, 1])\n",488    "ax.legend()\n",489    "plt.show()"490   ]491  },492  {493   "cell_type": "code",494   "execution_count": 14,495   "metadata": {496    "pycharm": {497     "is_executing": true498    }499   },500   "outputs": [501    {502     "data": {503      "image/png": "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\n",504      "text/plain": [505       "<Figure size 640x480 with 1 Axes>"506      ]507     },508     "metadata": {},509     "output_type": "display_data"510    },511    {512     "name": "stdout",513     "output_type": "stream",514     "text": [515      "time: 156 ms (started: 2023-04-05 21:18:45 +03:00)\n"516     ]517    }518   ],519   "source": [520    "plt.title('Loss')\n",521    "plt.plot(history.history['loss'], label='train')\n",522    "plt.plot(history.history['val_loss'], label='test')\n",523    "plt.legend()\n",524    "plt.show();"525   ]526  },527  {528   "cell_type": "code",529   "execution_count": 15,530   "metadata": {531    "pycharm": {532     "is_executing": true533    }534   },535   "outputs": [536    {537     "data": {538      "image/png": "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\n",539      "text/plain": [540       "<Figure size 640x480 with 1 Axes>"541      ]542     },543     "metadata": {},544     "output_type": "display_data"545    },546    {547     "name": "stdout",548     "output_type": "stream",549     "text": [550      "time: 125 ms (started: 2023-04-05 21:18:46 +03:00)\n"551     ]552    }553   ],554   "source": [555    "plt.title('Accuracy')\n",556    "plt.plot(history.history['accuracy'], label='train')\n",557    "plt.plot(history.history['val_accuracy'], label='test')\n",558    "plt.legend()\n",559    "plt.show();"560   ]561  },562  {563   "cell_type": "code",564   "execution_count": null,565   "metadata": {},566   "outputs": [],567   "source": [568    "#model.save(\"tf_model.h5\")"569   ]570  },571  {572   "cell_type": "code",573   "execution_count": null,574   "metadata": {},575   "outputs": [],576   "source": []577  },578  {579   "cell_type": "code",580   "execution_count": null,581   "metadata": {},582   "outputs": [],583   "source": []584  }585 ],586 "metadata": {587  "kernelspec": {588   "display_name": "Python 3 (ipykernel)",589   "language": "python",590   "name": "python3"591  },592  "language_info": {593   "codemirror_mode": {594    "name": "ipython",595    "version": 3596   },597   "file_extension": ".py",598   "mimetype": "text/x-python",599   "name": "python",600   "nbconvert_exporter": "python",601   "pygments_lexer": "ipython3",602   "version": "3.9.13"603  }604 },605 "nbformat": 4,606 "nbformat_minor": 1607}608