IMDB-GloVe-Bi-LSTM.ipynb

{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.7.12","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"# This Python 3 environment comes with many helpful analytics libraries installed\n# It is defined by the kaggle/python Docker image: https://github.com/kaggle/docker-python\n# For example, here's several helpful packages to load\n\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n\n# Input data files are available in the read-only \"../input/\" directory\n# For example, running this (by clicking run or pressing Shift+Enter) will list all files under the input directory\n\nimport os\nfor dirname, _, filenames in os.walk('/kaggle/input'):\n     for filename in filenames:\n        print(os.path.join(dirname, filename))\n# You can write up to 20GB to the current directory (/kaggle/working/) that gets preserved as output when you create a version using \"Save & Run All\" \n# You can also write temporary files to /kaggle/temp/, but they won't be saved outside of the current session","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","execution":{"iopub.status.busy":"2022-08-23T13:42:42.355730Z","iopub.execute_input":"2022-08-23T13:42:42.356497Z","iopub.status.idle":"2022-08-23T13:42:42.400487Z","shell.execute_reply.started":"2022-08-23T13:42:42.356343Z","shell.execute_reply":"2022-08-23T13:42:42.399149Z"},"trusted":true},"execution_count":1,"outputs":[{"name":"stdout","text":"/kaggle/input/glove-embeddings/glove.6B.200d.txt\n/kaggle/input/glove-embeddings/glove.6B.50d.txt\n/kaggle/input/glove-embeddings/glove.6B.300d.txt\n/kaggle/input/glove-embeddings/glove.6B.100d.txt\n/kaggle/input/imdb-spoiler-dataset/IMDB_movie_details.json\n/kaggle/input/imdb-spoiler-dataset/IMDB_reviews.json\n","output_type":"stream"}]},{"cell_type":"markdown","source":"## Important Library","metadata":{}},{"cell_type":"code","source":"# regular expression and string case\nimport re\nimport string\n\n# Data reading\nimport pandas as pd\nimport numpy as np\n\n# tensorflow and keras\nimport tensorflow as tf\nfrom tensorflow.keras.preprocessing.text import Tokenizer\nfrom tensorflow.keras.preprocessing.sequence import pad_sequences\nimport tensorflow_datasets as tfds\n\n# Sklearn\nfrom sklearn.preprocessing import LabelEncoder\nfrom sklearn.model_selection import train_test_split\n\n# plotting\nimport matplotlib.pyplot as plt\nimport seaborn as sns","metadata":{"execution":{"iopub.status.busy":"2022-08-23T13:42:42.402473Z","iopub.execute_input":"2022-08-23T13:42:42.402968Z","iopub.status.idle":"2022-08-23T13:42:50.751670Z","shell.execute_reply.started":"2022-08-23T13:42:42.402922Z","shell.execute_reply":"2022-08-23T13:42:50.750770Z"},"trusted":true},"execution_count":2,"outputs":[]},{"cell_type":"code","source":"# Read the dataset\ndf_reviews = pd.read_json('../input/imdb-spoiler-dataset/IMDB_reviews.json', lines=True)\ndf_details = pd.read_json('../input/imdb-spoiler-dataset/IMDB_movie_details.json',lines =True)","metadata":{"execution":{"iopub.status.busy":"2022-08-23T13:42:50.753202Z","iopub.execute_input":"2022-08-23T13:42:50.753546Z","iopub.status.idle":"2022-08-23T13:43:10.558960Z","shell.execute_reply.started":"2022-08-23T13:42:50.753494Z","shell.execute_reply":"2022-08-23T13:43:10.558064Z"},"trusted":true},"execution_count":3,"outputs":[]},{"cell_type":"markdown","source":"### Function to get a equal number of data from both classification  ","metadata":{}},{"cell_type":"code","source":"def get_part_data(df, num_reviews):\n    \n    num_pos_reviews = df[df[\"is_spoiler\"]== True].shape[0]  # Number of spoilers in the dataset\n    num_neg_reviews = df[df[\"is_spoiler\"]== False].shape[0] # Number of Non-spoilers in the dataset\n    \n    fraction_pos = num_reviews/num_pos_reviews  # fraction of spoiler reviews to be returned\n    fraction_neg = num_reviews/num_neg_reviews  # fraction of non-spoiler reviews to be returned \n\n    df_pos = df[df['is_spoiler'] == True].sample(frac = fraction_pos, random_state = 2)\n    df_neg = df[df['is_spoiler'] == False].sample(frac = fraction_neg, random_state = 2)\n\n    df_re = pd.concat([df_pos, df_neg])  # join the True and False dataset\n    df_re = df_re.reset_index(drop=True)  # mix the index values\n    df_re.loc[(df_re['is_spoiler'] == True) ,'is_spoiler'] =1\n    df_re.loc[(df_re['is_spoiler'] == False) ,'is_spoiler'] =0\n    return df_re[[\"movie_id\",\"review_text\",\"review_summary\",\"is_spoiler\"]]","metadata":{"execution":{"iopub.status.busy":"2022-08-23T13:43:10.561237Z","iopub.execute_input":"2022-08-23T13:43:10.561552Z","iopub.status.idle":"2022-08-23T13:43:10.569863Z","shell.execute_reply.started":"2022-08-23T13:43:10.561512Z","shell.execute_reply":"2022-08-23T13:43:10.568653Z"},"trusted":true},"execution_count":4,"outputs":[]},{"cell_type":"code","source":"old_data = get_part_data(df_reviews, 110000)","metadata":{"execution":{"iopub.status.busy":"2022-08-23T13:43:10.571538Z","iopub.execute_input":"2022-08-23T13:43:10.572111Z","iopub.status.idle":"2022-08-23T13:43:11.547621Z","shell.execute_reply.started":"2022-08-23T13:43:10.572069Z","shell.execute_reply":"2022-08-23T13:43:11.546771Z"},"trusted":true},"execution_count":5,"outputs":[]},{"cell_type":"markdown","source":"## Preprocess the data","metadata":{}},{"cell_type":"code","source":"stopwords = [ \"a\", \"about\", \"above\", \"after\", \"again\", \"against\", \"all\", \"am\", \"an\", \"and\", \"any\", \"are\", \"as\", \"at\", \"be\", \"because\", \n             \"been\", \"before\", \"being\", \"below\", \"between\", \"both\", \"but\", \"by\", \"could\", \"did\", \"do\", \"does\", \"doing\", \"down\", \"during\",\n             \"each\", \"few\", \"for\", \"from\", \"further\", \"had\", \"has\", \"have\", \"having\", \"he\", \"he'd\", \"he'll\", \"he's\", \"her\", \"here\", \n             \"here's\", \"hers\", \"herself\", \"him\", \"himself\", \"his\", \"how\", \"how's\", \"i\", \"i'd\", \"i'll\", \"i'm\", \"i've\", \"if\", \"in\", \"into\",\n             \"is\", \"it\", \"it's\", \"its\", \"itself\", \"let's\", \"me\", \"more\", \"most\", \"my\", \"myself\", \"nor\", \"of\", \"on\", \"once\", \"only\", \"or\",\n             \"other\", \"ought\", \"our\", \"ours\", \"ourselves\", \"out\", \"over\", \"own\", \"same\", \"she\", \"she'd\", \"she'll\", \"she's\", \"should\", \n             \"so\", \"some\", \"such\", \"than\", \"that\", \"that's\", \"the\", \"their\", \"theirs\", \"them\", \"themselves\", \"then\", \"there\", \"there's\",\n             \"these\", \"they\", \"they'd\", \"they'll\", \"they're\", \"they've\", \"this\", \"those\", \"through\", \"to\", \"too\", \"under\", \"until\", \"up\",\n             \"very\", \"was\", \"we\", \"we'd\", \"we'll\", \"we're\", \"we've\", \"were\", \"what\", \"what's\", \"when\", \"when's\", \"where\", \"where's\",\n             \"which\", \"while\", \"who\", \"who's\", \"whom\", \"why\", \"why's\", \"with\", \"would\", \"you\", \"you'd\", \"you'll\", \"you're\", \"you've\",\n             \"your\", \"yours\", \"yourself\", \"yourselves\" ]","metadata":{"execution":{"iopub.status.busy":"2022-08-23T13:43:11.548962Z","iopub.execute_input":"2022-08-23T13:43:11.549252Z","iopub.status.idle":"2022-08-23T13:43:11.560622Z","shell.execute_reply.started":"2022-08-23T13:43:11.549204Z","shell.execute_reply":"2022-08-23T13:43:11.559770Z"},"trusted":true},"execution_count":6,"outputs":[]},{"cell_type":"code","source":"# Removing the stop word and tags\ndef remove_stopwords(data):\n    data['review_ws'] = data['review_text'].apply(lambda x :' '.join([word for word in x.split() \n                                                                 if word not in (stopwords)]))\n    return data\n\ndef remove_tags(string):\n    result = re.sub('<.*?>','',string)\n    return result\n    \n# open new a feature as cleaned review text\ndata_ws = remove_stopwords(old_data)\ndata_ws['clean_review']= data_ws['review_ws'].apply(lambda cw : remove_tags(cw))\ndata_ws['clean_review'] = data_ws['clean_review'].str.replace('[{}]'.format(string.punctuation), ' ')","metadata":{"execution":{"iopub.status.busy":"2022-08-23T13:43:11.562201Z","iopub.execute_input":"2022-08-23T13:43:11.562718Z","iopub.status.idle":"2022-08-23T13:45:28.297588Z","shell.execute_reply.started":"2022-08-23T13:43:11.562675Z","shell.execute_reply":"2022-08-23T13:45:28.296665Z"},"trusted":true},"execution_count":7,"outputs":[{"name":"stderr","text":"/opt/conda/lib/python3.7/site-packages/ipykernel_launcher.py:14: FutureWarning: The default value of regex will change from True to False in a future version.\n  \n","output_type":"stream"}]},{"cell_type":"code","source":"# label encoder used to tranform the labels\nle = LabelEncoder()\n\n# train and test split\ntraining_reviews,testing_reviews,training_labels,testing_labels  = train_test_split(data_ws['clean_review'].values,old_data['is_spoiler'].values,test_size = 0.2)\ntraining_labels = le.fit_transform(training_labels)\ntesting_labels = le.fit_transform(testing_labels)","metadata":{"execution":{"iopub.status.busy":"2022-08-23T13:45:28.299130Z","iopub.execute_input":"2022-08-23T13:45:28.299551Z","iopub.status.idle":"2022-08-23T13:45:28.393460Z","shell.execute_reply.started":"2022-08-23T13:45:28.299500Z","shell.execute_reply":"2022-08-23T13:45:28.392543Z"},"trusted":true},"execution_count":8,"outputs":[]},{"cell_type":"code","source":"# Shape of train and test data\nprint(training_reviews.shape)\nprint(testing_reviews.shape)","metadata":{"execution":{"iopub.status.busy":"2022-08-23T13:45:28.394923Z","iopub.execute_input":"2022-08-23T13:45:28.395339Z","iopub.status.idle":"2022-08-23T13:45:28.402298Z","shell.execute_reply.started":"2022-08-23T13:45:28.395287Z","shell.execute_reply":"2022-08-23T13:45:28.401444Z"},"trusted":true},"execution_count":9,"outputs":[{"name":"stdout","text":"(176000,)\n(44000,)\n","output_type":"stream"}]},{"cell_type":"markdown","source":"### Tokenize the text\nUsing tokenizer to produce token for a given word and taking maximum length of 400 character of a review and after we simply truncate the input review and then padded the input to max len of 400","metadata":{}},{"cell_type":"code","source":"# Tokenize the text \ntokenizer = Tokenizer(num_words=10000,oov_token='<OOV>')\ntokenizer.fit_on_texts(training_reviews)\nword_index = tokenizer.word_index\ntraining_sequence = tokenizer.texts_to_sequences(training_reviews)\ntesting_sequence = tokenizer.texts_to_sequences(testing_reviews)\n\n# covert the text into padded sequence\ntrain_pad_sequence = pad_sequences(training_sequence,maxlen = 400,truncating= 'post',padding = 'pre')\ntest_pad_sequence = pad_sequences(testing_sequence,maxlen = 400,truncating= 'post',padding = 'pre')\nprint('Total Unique Words : {}'.format(len(word_index)))","metadata":{"execution":{"iopub.status.busy":"2022-08-23T13:45:28.405842Z","iopub.execute_input":"2022-08-23T13:45:28.406935Z","iopub.status.idle":"2022-08-23T13:46:38.960198Z","shell.execute_reply.started":"2022-08-23T13:45:28.406707Z","shell.execute_reply":"2022-08-23T13:46:38.959130Z"},"trusted":true},"execution_count":10,"outputs":[{"name":"stdout","text":"Total Unique Words : 171274\n","output_type":"stream"}]},{"cell_type":"markdown","source":"### Using Glove embedding with LSTM layer\n\nlayer1: Embedding Layer using glove weights\n\nlayer2: Using a Bidirectional LSTM\n\nlayer3: A dropout Layer\n\nlayer4: A Dense layer of 256 neurons with 'relu' activation\n\nlayer5: A Dense Layer of 128 neurons with 'relu' activation\n\nlayer6: Again a dropout layer.\n\nlayer7: Sigmoid activation layer to classify it positive and negative.","metadata":{}},{"cell_type":"code","source":"# Getting the Glove embedding text\nembedded_words = {}\nwith open('../input/glove-embeddings/glove.6B.300d.txt') as file:\n    for line in file:\n        words, coeff = line.split(maxsplit=1)\n        coeff = np.array(coeff.split(),dtype = float)\n        embedded_words[words] = coeff","metadata":{"execution":{"iopub.status.busy":"2022-08-23T13:46:38.961808Z","iopub.execute_input":"2022-08-23T13:46:38.962161Z","iopub.status.idle":"2022-08-23T13:47:10.402915Z","shell.execute_reply.started":"2022-08-23T13:46:38.962109Z","shell.execute_reply":"2022-08-23T13:47:10.402030Z"},"trusted":true},"execution_count":11,"outputs":[]},{"cell_type":"code","source":"# Check the word items using Glove embedding\nembedding_matrix = np.zeros((len(word_index) + 1,300))\nfor word, i in word_index.items():\n    embedding_vector = embedded_words.get(word)\n    if embedding_vector is not None:\n        embedding_matrix[i] = embedding_vector","metadata":{"execution":{"iopub.status.busy":"2022-08-23T13:47:10.404676Z","iopub.execute_input":"2022-08-23T13:47:10.405013Z","iopub.status.idle":"2022-08-23T13:47:10.868491Z","shell.execute_reply.started":"2022-08-23T13:47:10.404968Z","shell.execute_reply":"2022-08-23T13:47:10.867621Z"},"trusted":true},"execution_count":12,"outputs":[]},{"cell_type":"code","source":"# Built the Bi-LSTM model using tensorflow.keras\nmodel = tf.keras.Sequential([tf.keras.layers.Embedding(len(word_index) + 1,300,weights=[embedding_matrix],input_length=300,\n                            trainable=False),\n                             tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(64)),\n                             tf.keras.layers.Dropout(0.5),\n                             tf.keras.layers.Dense(256,activation = 'relu',),\n                             tf.keras.layers.Dense(128,activation = 'relu'),\n                             tf.keras.layers.Dropout(0.5),\n                             tf.keras.layers.Dense(1,activation = tf.nn.sigmoid)])","metadata":{"execution":{"iopub.status.busy":"2022-08-23T13:47:10.869855Z","iopub.execute_input":"2022-08-23T13:47:10.870260Z","iopub.status.idle":"2022-08-23T13:47:15.034845Z","shell.execute_reply.started":"2022-08-23T13:47:10.870204Z","shell.execute_reply":"2022-08-23T13:47:15.033960Z"},"trusted":true},"execution_count":13,"outputs":[{"name":"stderr","text":"2022-08-23 13:47:10.984018: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-08-23 13:47:11.099697: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-08-23 13:47:11.100773: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-08-23 13:47:11.102365: I tensorflow/core/platform/cpu_feature_guard.cc:142] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 AVX512F FMA\nTo enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n2022-08-23 13:47:11.103778: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-08-23 13:47:11.104708: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-08-23 13:47:11.105554: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-08-23 13:47:13.671352: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-08-23 13:47:13.672376: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-08-23 13:47:13.673217: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-08-23 13:47:13.673972: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1510] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 15403 MB memory:  -> device: 0, name: Tesla P100-PCIE-16GB, pci bus id: 0000:00:04.0, compute capability: 6.0\n","output_type":"stream"}]},{"cell_type":"code","source":"model.summary()","metadata":{"execution":{"iopub.status.busy":"2022-08-23T13:47:15.036188Z","iopub.execute_input":"2022-08-23T13:47:15.036520Z","iopub.status.idle":"2022-08-23T13:47:15.048129Z","shell.execute_reply.started":"2022-08-23T13:47:15.036476Z","shell.execute_reply":"2022-08-23T13:47:15.047104Z"},"trusted":true},"execution_count":14,"outputs":[{"name":"stdout","text":"Model: \"sequential\"\n_________________________________________________________________\nLayer (type)                 Output Shape              Param #   \n=================================================================\nembedding (Embedding)        (None, 300, 300)          51382500  \n_________________________________________________________________\nbidirectional (Bidirectional (None, 128)               186880    \n_________________________________________________________________\ndropout (Dropout)            (None, 128)               0         \n_________________________________________________________________\ndense (Dense)                (None, 256)               33024     \n_________________________________________________________________\ndense_1 (Dense)              (None, 128)               32896     \n_________________________________________________________________\ndropout_1 (Dropout)          (None, 128)               0         \n_________________________________________________________________\ndense_2 (Dense)              (None, 1)                 129       \n=================================================================\nTotal params: 51,635,429\nTrainable params: 252,929\nNon-trainable params: 51,382,500\n_________________________________________________________________\n","output_type":"stream"}]},{"cell_type":"code","source":"# Compile the loss using cross entropy function\nmodel.compile(loss = tf.keras.losses.BinaryCrossentropy() , optimizer='Adam' , metrics = 'accuracy')\nhistory = model.fit(train_pad_sequence,training_labels,epochs = 10 ,validation_data=(test_pad_sequence,testing_labels))","metadata":{"execution":{"iopub.status.busy":"2022-08-23T13:47:15.049553Z","iopub.execute_input":"2022-08-23T13:47:15.050062Z","iopub.status.idle":"2022-08-23T14:21:40.506507Z","shell.execute_reply.started":"2022-08-23T13:47:15.050016Z","shell.execute_reply":"2022-08-23T14:21:40.505536Z"},"trusted":true},"execution_count":15,"outputs":[{"name":"stderr","text":"2022-08-23 13:47:15.384494: I tensorflow/compiler/mlir/mlir_graph_optimization_pass.cc:185] None of the MLIR Optimization Passes are enabled (registered 2)\n","output_type":"stream"},{"name":"stdout","text":"Epoch 1/10\n","output_type":"stream"},{"name":"stderr","text":"2022-08-23 13:47:19.549839: I tensorflow/stream_executor/cuda/cuda_dnn.cc:369] Loaded cuDNN version 8005\n","output_type":"stream"},{"name":"stdout","text":"5500/5500 [==============================] - 213s 38ms/step - loss: 0.6382 - accuracy: 0.6253 - val_loss: 0.5870 - val_accuracy: 0.6863\nEpoch 2/10\n5500/5500 [==============================] - 205s 37ms/step - loss: 0.5858 - accuracy: 0.6896 - val_loss: 0.5743 - val_accuracy: 0.6931\nEpoch 3/10\n5500/5500 [==============================] - 207s 38ms/step - loss: 0.5700 - accuracy: 0.7019 - val_loss: 0.5675 - val_accuracy: 0.6974\nEpoch 4/10\n5500/5500 [==============================] - 206s 38ms/step - loss: 0.5592 - accuracy: 0.7093 - val_loss: 0.5676 - val_accuracy: 0.7009\nEpoch 5/10\n5500/5500 [==============================] - 205s 37ms/step - loss: 0.5476 - accuracy: 0.7191 - val_loss: 0.5647 - val_accuracy: 0.7015\nEpoch 6/10\n5500/5500 [==============================] - 204s 37ms/step - loss: 0.5364 - accuracy: 0.7270 - val_loss: 0.5680 - val_accuracy: 0.6987\nEpoch 7/10\n5500/5500 [==============================] - 204s 37ms/step - loss: 0.5250 - accuracy: 0.7364 - val_loss: 0.5689 - val_accuracy: 0.6999\nEpoch 8/10\n5500/5500 [==============================] - 206s 37ms/step - loss: 0.5130 - accuracy: 0.7442 - val_loss: 0.5722 - val_accuracy: 0.7024\nEpoch 9/10\n5500/5500 [==============================] - 207s 38ms/step - loss: 0.5001 - accuracy: 0.7528 - val_loss: 0.5765 - val_accuracy: 0.6962\nEpoch 10/10\n5500/5500 [==============================] - 207s 38ms/step - loss: 0.4877 - accuracy: 0.7612 - val_loss: 0.5803 - val_accuracy: 0.6973\n","output_type":"stream"}]},{"cell_type":"code","source":"# Evaluate the model using accuracy and validation accuracy\nacc = history.history['accuracy']\nval_acc = history.history['val_accuracy']\nloss = history.history['loss']\nval_loss = history.history['val_loss']\nepochs = range(len(acc))\n\nplt.plot(epochs, acc, 'r', label='Training accuracy')\nplt.plot(epochs, val_acc, 'b', label='Validation accuracy')\nplt.title('Training and validation accuracy')\nplt.legend(loc=0)\nplt.figure()\n\nplt.plot(epochs, loss, 'r', label='Training loss')\nplt.plot(epochs, val_loss, 'b', label='Validation loss')\nplt.title('Training and validation loss')\nplt.legend(loc=0)\n\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2022-08-23T14:21:40.510109Z","iopub.execute_input":"2022-08-23T14:21:40.510443Z","iopub.status.idle":"2022-08-23T14:21:41.021412Z","shell.execute_reply.started":"2022-08-23T14:21:40.510400Z","shell.execute_reply":"2022-08-23T14:21:41.020412Z"},"trusted":true},"execution_count":16,"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"}},{"output_type":"display_data","data":{"text/plain":"<Figure size 432x288 with 1 Axes>","image/png":"iVBORw0KGgoAAAANSUhEUgAAAXoAAAEICAYAAABRSj9aAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8/fFQqAAAACXBIWXMAAAsTAAALEwEAmpwYAAA0y0lEQVR4nO3de5xN9f7H8dfHDMkl5HJcZjSjhsjdoFLS5RTlcLpTYTgUJaVcu+BwOjqlkqJOKaqjQznlpyslldLFEGpcchsZ3URuufP5/fFdY/aMuWwze2bt2fN5Ph77Yfbaa6/1mY33fOe7vuv7FVXFGGNM5CrldwHGGGMKlwW9McZEOAt6Y4yJcBb0xhgT4SzojTEmwlnQG2NMhLOgNydFRN4TkV6h3tdPIpIqIpcVwnFVRM7yvn5WRB4MZt98nOdmEZmf3zpzOW4HEUkL9XFN0Yv2uwBT+ERkb8DTcsBB4Kj3/DZVnRHssVS1U2HsG+lUtX8ojiMiccAmoLSqHvGOPQMI+u/QlDwW9CWAqlZI/1pEUoG+qvph1v1EJDo9PIwxkcO6bkqw9F/NRWS4iPwMTBORKiLytohsE5Hfva9jAt7zsYj09b5OEpHPRGSCt+8mEemUz33jReRTEdkjIh+KyGQR+U8OdQdT4zgR+dw73nwRqRbweg8R2Swi20Xk/lw+n7Yi8rOIRAVsu1pEVnpftxGRL0Rkp4j8JCJPi0iZHI41XUT+EfB8qPeeH0WkT5Z9rxKRb0Rkt4hsEZExAS9/6v25U0T2ish56Z9twPvPF5ElIrLL+/P8YD+b3IhIQ+/9O0UkRUS6BLx2pYis8o65VUSGeNureX8/O0Vkh4gsEhHLnSJmH7ipCZwOnAHcivs3Mc17XhfYDzydy/vbAmuBasAjwAsiIvnY91Xga6AqMAbokcs5g6nxJqA3UAMoA6QHTyPgGe/4tb3zxZANVf0K+AO4JMtxX/W+PgoM9r6f84BLgdtzqRuvho5ePX8GEoCs1wf+AHoClYGrgAEi8lfvtfben5VVtYKqfpHl2KcD7wCTvO/tceAdEama5Xs44bPJo+bSwFvAfO99dwIzRKSBt8sLuG7AikBj4CNv+71AGlAd+BNwH2DzrhQxC3pzDBitqgdVdb+qblfV/6nqPlXdAzwEXJTL+zer6vOqehR4CaiF+w8d9L4iUhdoDYxS1UOq+hkwN6cTBlnjNFX9XlX3A68Bzb3t1wFvq+qnqnoQeND7DHLyX6A7gIhUBK70tqGqS1X1S1U9oqqpwL+zqSM7N3j1faeqf+B+sAV+fx+r6reqekxVV3rnC+a44H4wrFPVV7y6/gusAf4SsE9On01uzgUqAA97f0cfAW/jfTbAYaCRiJymqr+r6rKA7bWAM1T1sKouUptgq8hZ0Jttqnog/YmIlBORf3tdG7txXQWVA7svsvg5/QtV3ed9WeEk960N7AjYBrAlp4KDrPHngK/3BdRUO/DYXtBuz+lcuNb7NSJyCnANsExVN3t11Pe6JX726vgnrnWfl0w1AJuzfH9tRWSh1zW1C+gf5HHTj705y7bNQJ2A5zl9NnnWrKqBPxQDj3st7ofgZhH5RETO87Y/CqwH5ovIRhEZEdy3YULJgt5kbV3dCzQA2qrqaWR0FeTUHRMKPwGni0i5gG2xuexfkBp/Cjy2d86qOe2sqqtwgdaJzN024LqA1gAJXh335acGXPdToFdxv9HEqmol4NmA4+bVGv4R16UVqC6wNYi68jpubJb+9ePHVdUlqtoV160zB/ebAqq6R1XvVdV6QBfgHhG5tIC1mJNkQW+yqojr897p9feOLuwTei3kZGCMiJTxWoN/yeUtBalxNtBZRC7wLpyOJe//B68Cd+F+oLyepY7dwF4RORsYEGQNrwFJItLI+0GTtf6KuN9wDohIG9wPmHTbcF1N9XI49rtAfRG5SUSiReRGoBGum6UgvsK1/oeJSGkR6YD7O5rp/Z3dLCKVVPUw7jM5BiAinUXkLO9azC7cdY3cuspMIbCgN1lNBE4FfgO+BN4vovPejLuguR34BzALN94/OxPJZ42qmgLcgQvvn4DfcRcLc5PeR/6Rqv4WsH0ILoT3AM97NQdTw3ve9/ARrlvjoyy73A6MFZE9wCi81rH33n24axKfeyNZzs1y7O1AZ9xvPduBYUDnLHWfNFU9hAv2TrjPfQrQU1XXeLv0AFK9Lqz+uL9PcBebPwT2Al8AU1R1YUFqMSdP7LqICUciMgtYo6qF/huFMZHOWvQmLIhIaxE5U0RKecMPu+L6eo0xBWR3xppwURN4A3dhNA0YoKrf+FuSMZHBum6MMSbCWdeNMcZEuLDruqlWrZrGxcX5XYYxxhQrS5cu/U1Vq2f3WtgFfVxcHMnJyX6XYYwxxYqIZL0j+jjrujHGmAhnQW+MMRHOgt4YYyJcUH303g0sTwJRwFRVfTibfW7ATbeqwApVvSngtdOAVcAcVR0YgrqNMSF0+PBh0tLSOHDgQN47G1+VLVuWmJgYSpcuHfR78gx6b+rXybhFEtKAJSIy15vVL32fBGAk0E5VfxeRGlkOM46MlXGMMWEmLS2NihUrEhcXR87rxhi/qSrbt28nLS2N+Pj4oN8XTNdNG2C9qm70Jjaaibs9PVA/YLKq/u4V82v6CyLSCrcQRchXqTfGhMaBAweoWrWqhXyYExGqVq160r95BRP0dci8SEIamRcxAKiPmxr1cxH50uvqwZu7+jHyWKpMRG4VkWQRSd62bVvw1RtjQsZCvnjIz99TqC7GRuOmI+2AW1rseRGpjJtu9V1VzXUaWFV9TlUTVTWxevVsx/vnbedOePBBWLs2f+83xpgIFUzQbyXzajgxnLhaTRow11sTchPwPS74zwMGikgqMAHoKSInXMgNiUOH4LHH4J//LJTDG2MKz/bt22nevDnNmzenZs2a1KlT5/jzQ4cO5fre5ORkBg0alOc5zj///JDU+vHHH9O5c+eQHKuoBDPqZgmQICLxuIDvRuYVb8BNJ9sdmCYi1XBdORtVNX3xAUQkCUhU1cJZM7JGDejfHyZNci37s84qlNMYY0KvatWqLF++HIAxY8ZQoUIFhgzJ6PE9cuQI0dHZx1ViYiKJiYl5nmPx4sUhqbU4yrNFr6pHgIHAPGA18JqqpojIWBHp4u02D9guIquAhcBQb6WbojV0KERHw/jxRX5qY0xoJSUl0b9/f9q2bcuwYcP4+uuvOe+882jRogXnn38+a71u2sAW9pgxY+jTpw8dOnSgXr16TJo06fjxKlSocHz/Dh06cN1113H22Wdz8803kz6L77vvvsvZZ59Nq1atGDRoUJ4t9x07dvDXv/6Vpk2bcu6557Jy5UoAPvnkk+O/kbRo0YI9e/bw008/0b59e5o3b07jxo1ZtGhRyD+znAQ1jl5V38WtRRm4bVTA1wrc4z1yOsZ0YHp+igxarVpw663wzDOuVW+Toxlz8u6+G7zWdcg0bw4TJ57029LS0li8eDFRUVHs3r2bRYsWER0dzYcffsh9993H//73vxPes2bNGhYuXMiePXto0KABAwYMOGHM+TfffENKSgq1a9emXbt2fP755yQmJnLbbbfx6aefEh8fT/fu3fOsb/To0bRo0YI5c+bw0Ucf0bNnT5YvX86ECROYPHky7dq1Y+/evZQtW5bnnnuOK664gvvvv5+jR4+yb9++k/488ivy7owdNgxKlYKHC+dSgDGm6Fx//fVERUUBsGvXLq6//noaN27M4MGDSUlJyfY9V111FaeccgrVqlWjRo0a/PLLLyfs06ZNG2JiYihVqhTNmzcnNTWVNWvWUK9evePj04MJ+s8++4wePXoAcMkll7B9+3Z2795Nu3btuOeee5g0aRI7d+4kOjqa1q1bM23aNMaMGcO3335LxYoV8/uxnLSwm72ywGJioE8feOEFuP9+iI3N+z3GmAz5aHkXlvLlyx//+sEHH+Tiiy/mzTffJDU1lQ4dOmT7nlNOOeX411FRURw5ciRf+xTEiBEjuOqqq3j33Xdp164d8+bNo3379nz66ae88847JCUlcc8999CzZ8+QnjcnkdeiBxgxAlThkUf8rsQYEyK7du2iTh13C8/06dNDfvwGDRqwceNGUlNTAZg1a1ae77nwwguZMWMG4Pr+q1WrxmmnncaGDRto0qQJw4cPp3Xr1qxZs4bNmzfzpz/9iX79+tG3b1+WLVsW8u8hJ5EZ9GecAb16wfPPw48/+l2NMSYEhg0bxsiRI2nRokXIW+AAp556KlOmTKFjx460atWKihUrUqlSpVzfM2bMGJYuXUrTpk0ZMWIEL730EgATJ06kcePGNG3alNKlS9OpUyc+/vhjmjVrRosWLZg1axZ33XVXyL+HnITdmrGJiYkakoVHNmyABg3gzjvhiScKfjxjItjq1atp2LCh32X4bu/evVSoUAFV5Y477iAhIYHBgwf7XdYJsvv7EpGlqprtONPIbNEDnHkm3HIL/PvfkM3FGGOMyer555+nefPmnHPOOezatYvbbrvN75JCInKDHuC+++DgQXfHrDHG5GHw4MEsX76cVatWMWPGDMqVK+d3SSER2UFfvz506wZTpsBvv/ldjTHG+CKygx7cEMt9+6yf3hhTYkV+0DdqBNddB089BTt2+F2NMcYUucgPeoAHHoA9e+DJJ/2uxBhjilzJCPqmTeHqq13Q79rldzXGmCwuvvhi5s2bl2nbxIkTGTBgQI7v6dChA+lDsa+88kp27tx5wj5jxoxhwoQJuZ57zpw5rFp1fGVURo0axYcffngS1WcvnKYzLhlBD65Vv2uX68IxxoSV7t27M3PmzEzbZs6cGdR8M+BmnaxcuXK+zp016MeOHctll12Wr2OFq5IT9C1bQufO7qLsnj1+V2OMCXDdddfxzjvvHF9kJDU1lR9//JELL7yQAQMGkJiYyDnnnMPo0aOzfX9cXBy/eSPrHnroIerXr88FF1xwfCpjcGPkW7duTbNmzbj22mvZt28fixcvZu7cuQwdOpTmzZuzYcMGkpKSmD17NgALFiygRYsWNGnShD59+nDw4MHj5xs9ejQtW7akSZMmrFmzJtfvz+/pjCNvUrPcPPggtG3rhlsOH+53NcaEJT9mKT799NNp06YN7733Hl27dmXmzJnccMMNiAgPPfQQp59+OkePHuXSSy9l5cqVNG3aNNvjLF26lJkzZ7J8+XKOHDlCy5YtadWqFQDXXHMN/fr1A+CBBx7ghRde4M4776RLly507tyZ6667LtOxDhw4QFJSEgsWLKB+/fr07NmTZ555hrvvvhuAatWqsWzZMqZMmcKECROYOnVqjt+f39MZl5wWPUCbNnDFFTBhAvzxh9/VGGMCBHbfBHbbvPbaa7Rs2ZIWLVqQkpKSqZslq0WLFnH11VdTrlw5TjvtNLp06XL8te+++44LL7yQJk2aMGPGjBynOU63du1a4uPjqV+/PgC9evXi008/Pf76NddcA0CrVq2OT4SWE7+nMy5ZLXqAUaOgXTt49lm4916/qzEm7Pg1S3HXrl0ZPHgwy5YtY9++fbRq1YpNmzYxYcIElixZQpUqVUhKSuLAgQP5On5SUhJz5syhWbNmTJ8+nY8//rhA9aZPdVyQaY6LajrjoFr0ItJRRNaKyHoRyXbNVxG5QURWiUiKiLzqbWsuIl9421aKyI0FqjYUzj8fLr0UHn0U9u/3uxpjjKdChQpcfPHF9OnT53hrfvfu3ZQvX55KlSrxyy+/8N577+V6jPbt2zNnzhz279/Pnj17eOutt46/tmfPHmrVqsXhw4ePTy0MULFiRfZkc92uQYMGpKamsn79egBeeeUVLrroonx9b35PZ5xni15EooDJwJ+BNGCJiMxV1VUB+yQAI4F2qvq7iNTwXtoH9FTVdSJSG1gqIvNUdWeBKy+IUaPgoovcNMZBrB5vjCka3bt35+qrrz7ehZM+re/ZZ59NbGws7dq1y/X9LVu25MYbb6RZs2bUqFGD1q1bH39t3LhxtG3blurVq9O2bdvj4d6tWzf69evHpEmTjl+EBShbtizTpk3j+uuv58iRI7Ru3Zr+/fvn6/tKX8u2adOmlCtXLtN0xgsXLqRUqVKcc845dOrUiZkzZ/Loo49SunRpKlSowMsvv5yvcwbKc5piETkPGKOqV3jPRwKo6viAfR4BvlfVnK9GuP1WANep6rqc9gnZNMV5uegiWL/eTWdctmzhn8+YMGbTFBcvhTFNcR1gS8DzNG9boPpAfRH5XES+FJGOWQ8iIm2AMsCGbF67VUSSRSR527ZtQZQUAqNGuUVJpk0rmvMZY4xPQjXqJhpIADoA3YHnRaRy+osiUgt4BeitqseyvllVn1PVRFVNrF69eohKysMll7j++vHjwRu7a4wxkSiYoN8KBK6wHeNtC5QGzFXVw6q6CfgeF/yIyGnAO8D9qvplwUsOERE3rn7LFvD6y4wpycJttTmTvfz8PQUT9EuABBGJF5EyQDdgbpZ95uBa84hINVxXzkZv/zeBl1V1NuHmiiugdWvXqj982O9qjPFN2bJl2b59u4V9mFNVtm/fTtmTvK6Y56gbVT0iIgOBeUAU8KKqpojIWCBZVed6r10uIquAo8BQVd0uIrcA7YGqIpLkHTJJVZefVJWFRcT11f/lLzBjBiQl+V2RMb6IiYkhLS2NIrtGZvKtbNmyxMTEnNR7Indx8GCpQqtWbv6b1ashuuTdQ2aMKf5K5uLgwUrvq1+/HmbN8rsaY4wJOQt6gK5doUkT+Mc/4OhRv6sxxpiQsqAHKFXKzVe/Zg3MDr9rxsYYUxAW9OmuvRYaNnSt+mMnDPU3xphiy4I+XVSUa9V/9x3MmeN3NcYYEzIW9IFuvBESEmDcODcaxxhjIoAFfaCoKLj/fre8zttv+12NMcaEhAV9VjfdBPXqwdix1qo3xkQEC/qsSpeGkSMhORnef9/vaowxpsAs6LPTsyfUrWutemNMRLCgz06ZMq5V/+WXsGCB39UYY0yBWNDnpHdvqFPHteqNMaYYs6DPySmnwPDhsGgRfPKJ39UYY0y+WdDnpm9fqFnTWvXGmGLNgj43p54Kw4bBRx/B55/7XY0xxuSLBX1ebrsNqld3d8saY0wxZEGfl3LlYMgQmDcPvvrK72qMMeakBRX0ItJRRNaKyHoRGZHDPjeIyCoRSRGRVwO29xKRdd6jV6gKL1K33w5Vq1qr3hhTLOUZ9CISBUwGOgGNgO4i0ijLPgnASKCdqp4D3O1tPx0YDbQF2gCjRaRKKL+BIlGhAgweDO+8A8uW+V2NMcaclGBa9G2A9aq6UVUPATOBrln26QdMVtXfAVT1V2/7FcAHqrrDe+0DoGNoSi9iAwdC5crWqjfGFDvBBH0dYEvA8zRvW6D6QH0R+VxEvhSRjifxXkTkVhFJFpHksF2FvlIluPtuN1f9ypV+V2OMMUEL1cXYaCAB6AB0B54XkcrBvllVn1PVRFVNrF69eohKKgSDBkHFim4VKmOMKSaCCfqtQGzA8xhvW6A0YK6qHlbVTcD3uOAP5r3FR5UqLuxnz4aUFL+rMcaYoAQT9EuABBGJF5EyQDdgbpZ95uBa84hINVxXzkZgHnC5iFTxLsJe7m0rvgYPdkMuH3rI70qMMSYoeQa9qh4BBuICejXwmqqmiMhYEeni7TYP2C4iq4CFwFBV3a6qO4BxuB8WS4Cx3rbiq2pVuOMOmDUL1q71uxpjjMmTaJjNt56YmKjJycl+l5G7X3+FuDi4/np46SW/qzHGGERkqaomZvea3RmbHzVqwIABMGMGrF/vdzXGGJMrC/r8GjLELTs4frzflRhjTK4s6POrVi3o1w9efhlSU/2uxhhjcmRBXxDDhkGpUvDww35XYowxObKgL4iYGPjb3+DFF2HLlrz3N8YYH1jQF9Tw4aAKjzzidyXGGJMtC/qCOuMMSEqC55+HH3/0uxpjjDmBBX0ojBwJR47Ao4/6XYkxxpzAgj4U6tWDW26Bf/8bfvnF72qMMSYTC/pQue8+OHgQHnvM70qMMSaTiAr6BQvg8GGfTl6/PnTvDlOmwG+/+VSEMcacKGKCft06+POfISHB9aAcPOhDEfffD/v2wRNP+HByY4zJXsQE/VlnuSVda9WC/v3d88mT4cCBIiyiYUM30dlTT8Hq1UV4YmOMyVnEBL0IdOoEixfDBx+4ySUHDnTXSSdOdA3tIjFmDERHQ7Nmrt++yE5sjDHZi5igTycCl10Gn34KCxfC2We7tULi42HCBNi7t5ALaNgQ1qyBm25yE541agRvvVXIJzXGmJxFXNCnE4EOHeCjj1zoN2sGQ4e6wH/4YdizpxBPXqMGTJ8On3wCFSpAly7QtatNfmaM8UXEBn2gCy+E+fPhiy+gTRt3f9MZZ8C4cbBzZyGeuH17+OYbNz3Chx+61v3DD8OhQ4V4UmOMyaxEBH26c891F2yXLHHhP2qU68sfPRp2FNYCh6VLu18l1qxxFxFGjnS/XixcWEgnNMaYzIIKehHpKCJrRWS9iIzI5vUkEdkmIsu9R9+A1x4RkRQRWS0ik0REQvkN5EdiIvzf/8GyZXDppTB2rAv8++4rxCHwsbHwv/+5nzQHD8Ill0CPHvDzz4V0QmOMcfIMehGJAiYDnYBGQHcRaZTNrrNUtbn3mOq993ygHdAUaAy0Bi4KVfEF1aKFy96VK+HKK12vSlycm2a+0GYyuPJKSEmBBx+E115zV4snT4ajRwvphMaYcLZzJ3z9NfznPzBzZuGcI5gWfRtgvapuVNVDwEyga5DHV6AsUAY4BSgNhN1kME2auA84JQX++lc3i0F8vButUygTUp56qvs14ttvoXVrNw60bVvXp2SMiTj798N338Ebb7gGZZ8+cMEFbtxGlSruv3+PHvCvfxXO+UVVc99B5Dqgo6r29Z73ANqq6sCAfZKA8cA24HtgsKpu8V6bAPQFBHhaVe/P5hy3ArcC1K1bt9XmzZsL/p0VwLp18M9/wiuvuCHx/fq5aedjYgrhZKquZT94sOvG6d8fHnrI/e0bY4qNo0dh82b4/vsTHz/84P6rp6tVy82akvVRrx6UKZO/84vIUlVNzPa1EAV9VWCvqh4UkduAG1X1EhE5C3gSuNHb9QNgmKouyul8iYmJmpycfBLfXuHZuNENhZ8+3a0Y2KcPjBjhRuyE3O7d7qrwpElQtaob9N+jhxsnaowJC6rw668ZAb52bcbXGzZkHlB32mnQoMGJYZ6QABUrhr62ggb9ecAYVb3Cez4SQFXH57B/FLBDVSuJyFCgrKqO814bBRxQ1RyXYwqnoE+3ebP7leqFF+DYMejVyw2eOfPMQjjZ8uUwYAB8+aUbnjllCpxzTiGcyBiTk9273W/22bXOd+/O2K9MGTfdSnqIBwZ79epF204raNBH47pjLgW2AkuAm1Q1JWCfWqr6k/f11cBwVT1XRG4E+gEdcV037wMTVTXHW0XDMejTpaW5IfHPPefWGbnlFjdSp379EJ/o2DH3U2X4cHdn1733uou35cuH+ETGlFzHjrl7GFetclNTBYZ54GA4EfdbfHZdLXXrQlSUb99CJgUKeu8AVwITgSjgRVV9SETGAsmqOldExgNdgCPADmCAqq7xWvdTgPa4C7Pvq+o9uZ0rnIM+3U8/ucWknn3WjZTs1s1NXNkou7FIBbFtm+srevFFNzxz0iR3h6115xgTtGPH3G/lKSkZj/RwD5yKqkaN7MP8zDOhbFn/6g9WgYO+KBWHoE/3669uhM7kye4fzPXXwwMPuFE8IfX5564759tv4aqr3OyY8fEhPokxxVtgoK9alRHqWQO9Th3XKDvnnIxHw4ZQubJvpYeEBX0h++03NwX9U0+5npauXeG669zNWLVqhegkhw+7E4we7fqNHngAhgyBU04J0QmMKR6OHXOjWLJrof/xR8Z+tWtnBHl6sDdqVPwDPScW9EVkxw7XuzJ5csYdto0audk0L70ULroIKlUq4EnS0txQzNmz3e+VU6a4gxsTYbIGenorPbtAz66FXtJGKFvQF7Fjx2DFCjeP2YIFbvbM/fvdRZvWrTOC/7zzCtAgf/99d6PVhg1uCcPHHgvhrw/GFJ30QA/sbsku0GvVyr6FXtICPScW9D47eNCNlkwP/q+/djdXnHqqG0F56aUu/Js1c+P1g3bggBv3OX68+4kxbhzcfru7y8uYMHLggAvzTZvcSJf0x4YNLuCzC/TAVroFet4s6MPMrl1uqvoFC1z4r1rltlet6uY6Sw/+evWCHGCzbp1r3c+f7ybweeYZd0+1MUXk4EEX5IEhHhjqP/2Uef/Spd3QxPj4zK3zRo3g9NOLvPyIYEEf5n780S2Qkt7iT0tz2+PiMkL/kkvc8K8cqboZ2u6+2x2wXz/X0rf/NSYEDh2CLVuyD/HUVPdPLjBKoqIygjwuLvMjPt612sNl/HmksKAvRlTdDRvprf2FCzMWR2naNCP427d3i1edYM8et27tk0+6e7D79nVDM204psnF4cOugZFdiG/aBFu3nhjksbHZh3hcnLtAaj2IRcuCvhg7etTNm5/e2v/sM/drcnS0W0glPfjbtnW/Dh+3cqXrs3/zTXe166qrXPfOn/98khcCTDg7dsxd6A/msW9fxtd//OFa6OmhnpbmjpWuVCk3iV/WAE9/xMRYkIcbC/oIsn8/LF6cEfzJya6lVb68G76ZPqKncWMvz9PS3JwNzz3nJtk/6yy44w5ISorcAcVh6OBBdzNPaqq7RhNMGAcT3PldlVLE3TiUXYjHxbnWeqaGgwl7FvQRbMcO+PjjjK6e779322vUcOvjRke7HwTHjhxDf/wJTU1Ff9+JlormWO06aOwZaIWKqHL8cewYmZ7ntC3YfU891QVH3brukf51bCzUrBkZfbVHj7rujU2bsn9k7cPOTrly7rMqyCPYY5Qta7/YRRoL+hJkyxYX+gsWuLH84FpvgY9SB/Yh235Btm9H9ChSsQKlav4JqXY6UqqU26dUNu8LYlt2++zd6+r64Qf3daDoaNeyzPoDIPDPypX9n95H1d0ElzXAN250f/7wg+vnTifiujfi4098VKlyYvCecor/36Mp3izoTfZ27IBp09zdtRs3uqEQt97qHrVrh/x0qq7bIj30s/szLS1zYILrlgoM/qw/DGJjQzPp1N69mcM76yNwrDe4aWizC/L4eFdbfheQMCY/LOhN7o4dc3faTp4M773n+lKuucZdvL3ggiJtah475i4l5PSD4Icfsl/Pt3r13H8rqFkzYwWgnLpXsi4MX6HCiQFer15Gn3a2o56M8YkFvQnehg3uhqsXXnDjOps0cYF/881hMx/+wYOu5Z/bD4M9ezK/JzraBX3gP/cyZdw84zm1yqtWte4UU3xY0JuTt28f/Pe/8PTTbtWrSpWgd283xUJCgt/V5WnXrhN/AJQunblVXru2XZA0kcOC3uSfqhvPOXmymzHz8GG44go3RPPKKyNjyIwxESC3oLf2jMmdCLRrB6++6prGY8fCd99Bly5uTP4jj8D27X5XaYzJRVBBLyIdRWStiKwXkRHZvJ4kIttEZLn36BvwWl0RmS8iq0VklYjEhbB+U5Rq1nRr127aBK+/7q5IDh/uxhH27g1Ll/pdoTEmG3kGvbfu62SgE9AI6C4i2a2OOktVm3uPqQHbXwYeVdWGQBvg1xDUbfxUurRbQmvhQre8Ye/eLvgTE928DP/5j7tiaowJC8G06NsA61V1o6oeAmYCXYM5uPcDIVpVPwBQ1b2qui+Pt5nipHFjNw5/61a3vNbvv0OPHm5c4/33u6ugxhhfBRP0dYDA/61p3rasrhWRlSIyW0RivW31gZ0i8oaIfCMij3q/IWQiIreKSLKIJG/btu2kvwkTBipVgjvvdMsCzZ8P558PDz/suneuvdZd0DXG+CJUF2PfAuJUtSnwAfCStz0auBAYArQG6gFJWd+sqs+paqKqJlavXj1EJRlflCrlZsicM8fdYjp0qOviadfOhf8bb7gB7caYIhNM0G8FYgOex3jbjlPV7aqa3ik7FWjlfZ0GLPe6fY4Ac4CWBarYFB9nnOFa9Vu2wFNPuVtar70WGjRwwzWzzilgjCkUwQT9EiBBROJFpAzQDZgbuIOIBK5K3QVYHfDeyiKS3ky/BFhVsJJNsVO+vLu79vvv3Vj8atXc87p13Sie7OY0MMaETJ5B77XEBwLzcAH+mqqmiMhYEeni7TZIRFJEZAUwCK97RlWP4rptFojIt4AAz4f+2zDFQlSUa9F/8YVbQeWii+Chh1zg9+2bsXiuMSak7M5Y46916+CJJ9wsmgcOuLtthwyBDh1sohljToLdGWvCV0KCG565ZYu763bJErcSemKim2sn65zFxpiTZkFvwkO1aq6/fvNmt+zhvn1w001w5pnw+OOwe7ffFRpTbFnQm/By6qnQrx+kpMBbb7mpJu+9192ANXSo3YBlTD5Y0JvwVKoUdO7sFsRdssT13T/xhAv+Hj3c1MnGmKBY0Jvwl95fv369G5b55pvQogVcdplbGSvMBhQYE24s6E3xERfnWvVpafCvf7npFjp1gqZNYfp0m0jNmBxY0Jvip3JlGDbMTZf80kuum6d3b/eDYPx4N7GaMeY4C3pTfJUpAz17uv76+fNdy/6++9yF20GD3Fw7xhgLehMBRNxEavPmwYoVbq78Z591Y/RvuAG++srvCo3xlQW9iSzp/fWbNrnhmPPnu8VQLrzQzahpM2eaEsiC3kSmOnUyZs6cONH9efXVrpX/+OOwc6ffFRpTZCzoTWSrWBHuussNzZw9261ve++97s/0GTWNiXAW9KZkiI52M2d++qlbxPzaa+H5593c+Fde6bp4bDy+iVAW9KbkadnSDcv84QcYMwaWLYMrroBzznEXcW1BFBNhLOhNyfWnP8Ho0W4itZdfhnLlYMAA160zbJjbbkwEsKA35pRT3Pw5S5a4BVH+/Gd3wbZePTdU87PPrFvHFGsW9MakE3GLmL/2mrvZasgQ+OgjNzQzMdG1+m2aBVMMBRX0ItJRRNaKyHoRGZHN60kisk1ElnuPvlleP01E0kTk6VAVbkyhqlvXzaezZYvrtz9wAHr1cguejxkDP//sd4XGBC3PoBeRKGAy0AloBHQXkUbZ7DpLVZt7j6lZXhsHfFrgao0pauXLw223wXffuZE5iYnw97+7HwQ9e7oRPMaEuWBa9G2A9aq6UVUPATOBrsGeQERaAX8C5uevRGPCQPo0C2+/7cbe9+/vpktOTHRdO7Nnw5EjfldpTLaCCfo6QOCyPmnetqyuFZGVIjJbRGIBRKQU8BgwJLcTiMitIpIsIsnbtm0LsnRjfJKQAJMmuemSn3gCtm6F6693yx4+8gjs2OF3hcZkEqqLsW8BcaraFPgAeMnbfjvwrqqm5fZmVX1OVRNVNbF69eohKsmYQlapEtx9N6xb5+bROfNMGD7cDc/s3x9WrfK7QmOA4IJ+KxAb8DzG23acqm5X1fThCFOBVt7X5wEDRSQVmAD0FJGHC1SxMeEmKgq6dnUjdFascIuaT5/ubsC64gp49104dszvKk0JFkzQLwESRCReRMoA3YC5gTuISK2Ap12A1QCqerOq1lXVOFz3zcuqesKoHWMiRtOmMHWq69Z56CF3Efeqq6BhQ3j6adi71+8KTQmUZ9Cr6hFgIDAPF+CvqWqKiIwVkS7eboNEJEVEVgCDgKTCKtiYYqFaNbcISmqqW++2ShW48043q+a997rtxhQR0TC74y8xMVGTk5P9LsOY0PvqK3jySXj9ddeVc801MHgwnHeeG9VjTAGIyFJVTczuNbsz1pii0rYtvPpqxqIoCxa4O3HPPRdmzoTDh/2u0EQoC3pjilpMTMaiKJMnu8XMu3d3c+s88ogtbm5CzoLeGL+ULw+33w5r1sBbb0H9+m54ZmysWxRl3Tq/KzQRwoLeGL+VKgWdO7uunOXL3c1X6YuidOkCCxfa7JmmQCzojQknzZrBtGluLvwHH4Qvv4RLLslYLMVmzzT5YEFvTDiqWdNNnvbDD25c/uHDkJQEcXEwbhzYVCHmJFjQGxPOypaFv/0Nvv3WzZ7ZogWMGuVmz+zXD1JS/K7QFAMW9MYUB+mzZ777rptDp1cvmDEDGjd20yy8/77145scWdAbU9w0bOgWQ9myxU2z8O230KmTm1vnuedg/36/KzRhxoLemOKqatWMaRZeeQVOPdUtkhIbC/ffDz/+6HeFJkxY0BtT3JUpA7fcAsnJ8MknbiGU8ePdhdsePWDZMr8rND6zoDcmUohA+/Zu5at162DAADdPfqtWcNFF7uujR/2u0vjAgt6YSHTmmW4CtbQ0mDDBde9cfbW7CWvSJNizx+8KTRGyoDcmklWq5KZF3rABXnsNatSAu+5y/fgjR8LPP/tdoSkCFvTGlATR0W5qhcWL3d22l1/uJlCLi3NdPBs2+F2hKUQW9MaUNG3butb9mjVuPP6LL7oJ1bp1g2++8bs6Uwgs6I0pqRIS4N//dv33Q4e6m7FatoSOHeHjj+0GrAgSVNCLSEcRWSsi60XkhDVfRSRJRLaJyHLv0dfb3lxEvvCWGVwpIjeG+hswxhRQrVpufvwffnDDMpcvh4svditfzZljC5tHgDyDXkSigMlAJ6AR0F1EGmWz6yxVbe49pnrb9gE9VfUcoCMwUUQqh6Z0Y0xIVa4MI0a4Fv6zz7qJ066+2t1xO20aHDrkd4Umn4Jp0bcB1qvqRlU9BMwEugZzcFX9XlXXeV//CPwKVM9vscaYIlC2rLvDdu1at8Rh2bLQp48bsvnEE7B3r98VmpMUTNDXAbYEPE/ztmV1rdc9M1tEYrO+KCJtgDLACZf3ReRWEUkWkeRtNv2qMeEhOhpuvNHdWfv++3DWWXDPPW7mzNGj4bff/K7QBClUF2PfAuJUtSnwAfBS4IsiUgt4Beitqid0+Knqc6qaqKqJ1atbg9+YsCLiZshcuBC++MLdZTt2rAv8QYPcIikmrAUT9FuBwBZ6jLftOFXdrqrpS99MBVqlvyYipwHvAPer6pcFK9cY46tzz3VTLKxa5YZjPvOM69Lp2RO++87v6kwOggn6JUCCiMSLSBmgGzA3cAevxZ6uC7Da214GeBN4WVVnh6ZkY4zvGjZ04+83bnSt+jfegCZN3Bq3ixf7XZ3JIs+gV9UjwEBgHi7AX1PVFBEZKyJdvN0GeUMoVwCDgCRv+w1AeyApYOhl81B/E8YYn8TGwuOPu6GZY8e6kG/Xzk2u9s47NhY/TIiG2V9EYmKiJicn+12GMSY//vjDtfQnTHDh36QJDB/uLupGR/tdXUQTkaWqmpjda3ZnrDEmdMqXhzvvhPXr4eWX3c1Wt9zi7sKdPBn27fO7whLJgt4YE3qlS7tFT1auhLfegtq1YeBAN4naP/4Bv//ud4UligW9MabwlCoFnTvD55/DokXQpg08+KAbmjlkCGzdmvcxTIFZ0BtjisYFF8Dbb8OKFdC1K0ycCPXqwe23u4XOTaGxoDfGFK2mTeE//3H9+H36wNSpbiz+gAHuAq4JOQt6Y4w/4uLcDVfr10PfvvDCC26aBQv8kLOgN8b4q25dmDIlI/BffNEFfv/+Nr1CiFjQG2PCQ2Dg9+vnpkZOSHAzaVrgF4gFvTEmvMTGujH3GzbArbfC9OkW+AVkQW+MCU8xMfD005kD/6yz3NepqX5XV6xY0Btjwltg4PfvDy+95Fr4/frBpk1+V1csWNAbY4qHmBh46qmMwH/lFahf3wI/CBb0xpjiJTDwBwzICPy+fd20yeYEFvTGmOKpTh2YNMmF++23u5uw6teHv/3NAj8LC3pjTPFWuzY8+aQL9zvugBkzLPCzsKA3xkSGwMAfODAj8Pv0cd08JZgFvTEmstSu7SZMSw/8//4XGjSA3r3dzVglkAW9MSYyBQb+nXfCzJlw9tklMvCDCnoR6Sgia0VkvYiMyOb1JBHZFrAubN+A13qJyDrv0SuUxRtjTJ5q1YInnshYyDw98JOSSkzg5xn0IhIFTAY6AY2A7iLSKJtdZ6lqc+8x1Xvv6cBooC3QBhgtIlVCVr0xxgSrVi23kPmmTS7wZ81ygd+rF6xb53d1hSqYFn0bYL2qblTVQ8BMoGuQx78C+EBVd6jq78AHQMf8lWqMMSFQs2ZG4N91F7z+ugv8m2+GlBS/qysUwQR9HSBw+Zc0b1tW14rIShGZLSKxJ/NeEblVRJJFJHnbtm1Blm6MMQVQsyY89pjr0rn3Xvi//4PGjeGaa2DpUr+rC6lQXYx9C4hT1aa4VvtLJ/NmVX1OVRNVNbF69eohKskYY4JQsyY88oibGfPBB+GjjyAxEa68EhYv9ru6kAgm6LcCsQHPY7xtx6nqdlU96D2dCrQK9r3GGBMWqlaFsWNd4P/zn7BkCbRrB5dcAgsWgKrfFeZbMEG/BEgQkXgRKQN0A+YG7iAitQKedgFWe1/PAy4XkSreRdjLvW3GGBOeKlWCkSPdVMiPPw5r1sBll8H557vFzYth4OcZ9Kp6BBiIC+jVwGuqmiIiY0Wki7fbIBFJEZEVwCAgyXvvDmAc7ofFEmCst80YY8Jb+fIweLDrw3/mGfjpJ/jLX6BlS5g9G44d87vCoImG2U+nxMRETU5O9rsMY4zJ7PBhN63C+PHw/ffQsCHcdx906wbR0X5Xh4gsVdXE7F6zO2ONMSYYpUu7m6xWrXI3XUVFQY8ebnqF55+HgwfzPIRfLOiNMeZkREXBjTfCihUwZw6cfrpb3vCss9w8+fv3+13hCSzojTEmP0qVgq5d4euvYd48iItzd9zGxbnhmnv2+F3hcRb0xhhTECJw+eWwaBF88gk0awbDh8MZZ7jhmr//7neFFvTGGBMy7dvD/Pnw1Vdw4YUwerQL/JEj4ddffSvLgt4YY0KtTRs3pcKKFe4O23/9y3Xp3H03bC36e0Yt6I0xprA0bepG6KxaBTfcAE8/DfXqQf/+blK1ImJBb4wxhe3ss2H6dDcdcu/eMG0aJCS4KZLXrCn001vQG2NMUYmPh2efzVj16vXXoVGjjOGahcSC3hhjilqdOm7Vq9RUN0LnvfegeXPXvVMIsxVY0BtjjF9q1HBTKmzeDH//u+vOEQn5afyfoMEYY0q6KlVg1KhCO7y16I0xJsJZ0BtjTISzoDfGmAhnQW+MMRHOgt4YYyKcBb0xxkQ4C3pjjIlwFvTGGBPhwm5xcBHZBmwuwCGqAb+FqJzizj6LzOzzyMw+jwyR8FmcoarVs3sh7IK+oEQkOaeV0Esa+ywys88jM/s8MkT6Z2FdN8YYE+Es6I0xJsJFYtA/53cBYcQ+i8zs88jMPo8MEf1ZRFwfvTHGmMwisUVvjDEmgAW9McZEuIgJehHpKCJrRWS9iIzwux4/iUisiCwUkVUikiIid/ldk99EJEpEvhGRt/2uxW8iUllEZovIGhFZLSLn+V2Tn0RksPf/5DsR+a+IlPW7plCLiKAXkShgMtAJaAR0F5FG/lblqyPAvaraCDgXuKOEfx4AdwGr/S4iTDwJvK+qZwPNKMGfi4jUAQYBiaraGIgCuvlbVehFRNADbYD1qrpRVQ8BM4GuPtfkG1X9SVWXeV/vwf1HruNvVf4RkRjgKmCq37X4TUQqAe2BFwBU9ZCq7vS1KP9FA6eKSDRQDvjR53pCLlKCvg6wJeB5GiU42AKJSBzQAvjK51L8NBEYBhzzuY5wEA9sA6Z5XVlTRaS830X5RVW3AhOAH4CfgF2qOt/fqkIvUoLeZENEKgD/A+5W1d1+1+MHEekM/KqqS/2uJUxEAy2BZ1S1BfAHUGKvaYlIFdxv//FAbaC8iNzib1WhFylBvxWIDXge420rsUSkNC7kZ6jqG37X46N2QBcRScV16V0iIv/xtyRfpQFpqpr+G95sXPCXVJcBm1R1m6oeBt4Azve5ppCLlKBfAiSISLyIlMFdTJnrc02+ERHB9cGuVtXH/a7HT6o6UlVjVDUO9+/iI1WNuBZbsFT1Z2CLiDTwNl0KrPKxJL/9AJwrIuW8/zeXEoEXp6P9LiAUVPWIiAwE5uGumr+oqik+l+WndkAP4FsRWe5tu09V3/WvJBNG7gRmeI2ijUBvn+vxjap+JSKzgWW40WrfEIHTIdgUCMYYE+EipevGGGNMDizojTEmwlnQG2NMhLOgN8aYCGdBb4wxEc6C3hhjIpwFvTHGRLj/BzKz20leu5LvAAAAAElFTkSuQmCC\n"},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","source":"print('Training Accuracy: {}'.format(max(acc)))\nprint('Validation Accuracy: {}'.format(max(val_acc)))","metadata":{"execution":{"iopub.status.busy":"2022-08-23T14:21:41.022971Z","iopub.execute_input":"2022-08-23T14:21:41.023518Z","iopub.status.idle":"2022-08-23T14:21:41.030622Z","shell.execute_reply.started":"2022-08-23T14:21:41.023467Z","shell.execute_reply":"2022-08-23T14:21:41.029468Z"},"trusted":true},"execution_count":17,"outputs":[{"name":"stdout","text":"Training Accuracy: 0.7611931562423706\nValidation Accuracy: 0.7024317979812622\n","output_type":"stream"}]},{"cell_type":"code","source":"\npredictions = (model.predict(test_pad_sequence) > 0.5).astype(\"int32\")","metadata":{"execution":{"iopub.status.busy":"2022-08-23T14:21:41.032491Z","iopub.execute_input":"2022-08-23T14:21:41.033486Z","iopub.status.idle":"2022-08-23T14:22:02.214805Z","shell.execute_reply.started":"2022-08-23T14:21:41.033426Z","shell.execute_reply":"2022-08-23T14:22:02.213857Z"},"trusted":true},"execution_count":18,"outputs":[]},{"cell_type":"markdown","source":"### Evaluation Report","metadata":{}},{"cell_type":"code","source":"from sklearn.metrics import classification_report\nreport = classification_report(testing_labels,predictions)\nprint(report)","metadata":{"execution":{"iopub.status.busy":"2022-08-23T14:22:02.216579Z","iopub.execute_input":"2022-08-23T14:22:02.216904Z","iopub.status.idle":"2022-08-23T14:22:02.297128Z","shell.execute_reply.started":"2022-08-23T14:22:02.216860Z","shell.execute_reply":"2022-08-23T14:22:02.296241Z"},"trusted":true},"execution_count":19,"outputs":[{"name":"stdout","text":"              precision    recall  f1-score   support\n\n           0       0.68      0.73      0.71     21917\n           1       0.71      0.66      0.69     22083\n\n    accuracy                           0.70     44000\n   macro avg       0.70      0.70      0.70     44000\nweighted avg       0.70      0.70      0.70     44000\n\n","output_type":"stream"}]},{"cell_type":"markdown","source":"### AUC score for the Glove embedding using BI-LStm Model","metadata":{}},{"cell_type":"code","source":"from sklearn.metrics import roc_auc_score\nprint(roc_auc_score(testing_labels, predictions ))","metadata":{"execution":{"iopub.status.busy":"2022-08-23T14:22:02.298725Z","iopub.execute_input":"2022-08-23T14:22:02.299225Z","iopub.status.idle":"2022-08-23T14:22:02.320046Z","shell.execute_reply.started":"2022-08-23T14:22:02.299173Z","shell.execute_reply":"2022-08-23T14:22:02.319104Z"},"trusted":true},"execution_count":20,"outputs":[{"name":"stdout","text":"0.697414013399046\n","output_type":"stream"}]},{"cell_type":"code","source":"import sklearn.metrics as metrics","metadata":{"execution":{"iopub.status.busy":"2022-08-23T14:25:01.288039Z","iopub.execute_input":"2022-08-23T14:25:01.288346Z","iopub.status.idle":"2022-08-23T14:25:01.294808Z","shell.execute_reply.started":"2022-08-23T14:25:01.288307Z","shell.execute_reply":"2022-08-23T14:25:01.293682Z"},"trusted":true},"execution_count":22,"outputs":[]},{"cell_type":"code","source":"fpr, tpr, _ = metrics.roc_curve(testing_labels,  predictions)\nauc = metrics.roc_auc_score(testing_labels, predictions)\nplt.plot(fpr,tpr,label=\"Pretrained Glove Bi-LSTM, auc=\"+str(auc))\nplt.legend(loc=4)\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2022-08-23T14:33:16.800487Z","iopub.execute_input":"2022-08-23T14:33:16.800789Z","iopub.status.idle":"2022-08-23T14:33:17.040151Z","shell.execute_reply.started":"2022-08-23T14:33:16.800756Z","shell.execute_reply":"2022-08-23T14:33:17.039278Z"},"trusted":true},"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"}}]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]}]}