Logistic Regression/Logistic_model_creation.ipynb

{"cells":[{"cell_type":"markdown","source":["Logistic regression is a statistical machine learning algorithm that classifies the data by considering outcome variables on extreme ends and tries makes a logarithmic line that distinguishes between them."],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":false,"cellMetadata":{},"nuid":"255f123a-7668-43d7-bcaf-7e5c0786914a","inputWidgets":{},"title":""}}},{"cell_type":"code","source":["from sklearn import datasets\nfrom sklearn import model_selection\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.metrics import confusion_matrix, classification_report ,plot_confusion_matrix\nimport numpy as np\nimport math\nimport matplotlib.pyplot as plt"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":true,"cellMetadata":{},"nuid":"63bb39f3-6f3d-48c8-978c-ce809b82cc53","inputWidgets":{},"title":"Importing the Required Libraries"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"code","source":["X,y=datasets.make_classification(n_samples=50000, n_features=15, n_informative=10, n_redundant=5, n_classes=2,weights=[0.7], class_sep=0.7,  random_state=15)\n"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":true,"cellMetadata":{},"nuid":"da7b404b-fb4b-45ea-8477-21a5792e6c3d","inputWidgets":{},"title":"Creating The Sample Dataset"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"code","source":["X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=15)"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":true,"cellMetadata":{},"nuid":"33367295-f0f7-4e0b-b8b9-25c91fe76c1e","inputWidgets":{},"title":"Using the Train Test Spliit"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"code","source":["def plot_confusion_matrix(cm, title='CONFUSION MATRIX', cmap=plt.cm.Reds):\n    target_names=['1','0']\n    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n    plt.title(title)\n    plt.colorbar()\n    tick_marks = np.arange(len(target_names))\n    plt.xticks(tick_marks, target_names, rotation=45)\n    plt.yticks(tick_marks, target_names)\n    plt.tight_layout()\n    plt.ylabel('True label')\n    plt.xlabel('Predicted label')"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":true,"cellMetadata":{},"nuid":"c82afaaa-460c-4bff-a94c-3fc8233a5db6","inputWidgets":{},"title":"Confusion Matrix Function"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"code","source":["def initialize_weights(shape):\n  \"\"\"\n  This function initializes the weights and bias of a model with a given shape. \n  The shape is specified as a tuple of integers, which represents the size of the weight matrix.\n  The function creates a weight matrix w with the same shape as the input data and fills it with zeros.\n  It also creates a bias term b and initializes it to 0. The weight matrix and bias are returned as a tuple.\n  \n  Here is a breakdown of each part of the function:\n  shape: This is a tuple of integers that specifies the size of the weight matrix.\n  b: This is the bias term, which is initialized to 0.\n  w: This is the weight matrix, which is initialized with the same shape as the input data and filled with zeros.\n  np.zeros_like(X_train[0]): This creates a weight matrix with the same shape as the first sample in the input data X_train.\n  The matrix is filled with zeros.\n  return w,b: This returns the weight matrix and bias as a tuple.\n  \"\"\"\n  \n  b = 0\n  w = np.zeros_like(X_train[0])\n  return w,b"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":true,"cellMetadata":{},"nuid":"f7e057b0-2337-440c-a11a-d27d5ddc60bd","inputWidgets":{},"title":"Initializing the weights"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"code","source":["def sigmoid(z):\n\t\"\"\"\n\tThis function calculates the sigmoid of a given input value z. The sigmoid function is defined as:\n    sigmoid(z) = 1 / (1 + exp(-z))\n    Where exp is the exponential function. The sigmoid function maps any input value to a value between 0 and 1. \n    It is often used as the activation function in neural networks and logistic regression models.\n\n    Here is a breakdown of each part of the function:\n    z: This is the input value for which the sigmoid function is being calculated.\n    sigmoid_: This is the output of the sigmoid function, which is calculated as 1/(1 + exp(-z)).\n    return sigmoid_: This returns the output of the sigmoid function.\n\t\"\"\"\n\tsigmoid_ = 1/(1+math.exp(-z)) # Your code goes here\n\treturn sigmoid_"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":true,"cellMetadata":{},"nuid":"4e73d54e-37a1-415e-a280-2f0e03c6e3c6","inputWidgets":{},"title":"Sigmoid Function"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"code","source":["def logloss(y_actual,y_pred):\n  \"\"\"\n  this function calculates the log loss of a set of predicted values y_pred given the actual values y_actual.\n  The log loss is a measure of the performance of a classification model, \n  where the prediction outcome is a probability between 0 and 1. It is defined as:\n\n  log loss = -(1/N) * ∑(y_actual * log(y_pred) + (1 - y_actual) * log(1 - y_pred))\n  \n  y_actual: This is a list or array of actual target values.\n\n  y_pred: This is a list or array of predicted target values, where each value is a probability between 0 and 1.\n\n  N: This is the length of the y_actual list. It represents the number of samples in the data set.\n\n  sum1: This variable is used to store the sum of the log loss for each sample in the data set. It is initialized to 0.\n\n   The for loop iterates over each sample in the data set. For each sample, the log loss is calculated as:\n   log loss = y_actual * log(y_pred) + (1 - y_actual) * log(1 - y_pred)\n   This value is then added to the sum1 variable.\n   Finally, the average log loss is calculated by dividing sum1 by N and multiplying the result by -1. \n   This final value is returned   as the output of the function.\n  \"\"\"\n  N=len(y_actual)\n  sum1=0\n  for i in range(N):\n    diff=1-y_pred[i]\n    s=(y_actual[i]*np.log10(y_pred[i])) + ((1-y_actual[i])*np.log10(diff))\n    sum1+=s\n  loss = -1 * 1/N *sum1 \n  return loss"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":true,"cellMetadata":{},"nuid":"3a1662e4-784c-47ea-9575-ed481ba0dc23","inputWidgets":{},"title":"Log Loss Function"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"code","source":["def gradient_wrtb(X,y,w,b):\n  \"\"\"\n  This function calculates the gradient of the log loss function with respect to the bias term b.\n  The log loss function is a measure of the performance of a classification model, \n  where the prediction outcome is a probability between 0 and 1.\n  The gradient of the log loss with respect to the bias is given by:\n  gradient(b) = sigmoid(w^T * X + b) - y\n  Where sigmoid is the sigmoid function, w is the weight matrix, X is the input data, \n  b is the bias term, and y is the actual target value.\n\n The function takes in the input data X, the actual target values y, the weight matrix w, and the bias term b as arguments.\n It calculates the gradient of the log loss with respect to the bias and returns the result.\n\n Here is a breakdown of each part of the function:\n X: This is the input data, which is an array of independent variables.\n y: This is the actual target values, which is an array of the same size as X.\n w: This is the weight matrix, which is an array of the same size as X.\n b: This is the bias term, which is a scalar value.\n gradient_db: This is the gradient of the log loss with respect to the bias term. It is calculated as sigmoid(np.dot(w.T, X )+ b)-y.\n return gradient_db: This returns the gradient of the log loss with respect to the bias term.\n \"\"\"\n  gradient_db = sigmoid(np.dot(w.T, X )+ b)-y\n  return gradient_db"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":true,"cellMetadata":{},"nuid":"9194e9d3-53ae-4a10-8975-8b34695ff825","inputWidgets":{},"title":"Gradient descent"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"code","source":["def gradient_wrtw(X,y,w,b):\n  \"\"\"This function calculates the gradient of the log loss function with respect to the weight matrix w.\n  The log loss function is a measure of the performance of a classification model, \n  where the prediction outcome is a probability between 0 and 1.\n  The gradient of the log loss with respect to the weight matrix is given by:\n  gradient(w) = X^T * (sigmoid(w^T * X + b) - y)\n  Where sigmoid is the sigmoid function, X is the input data, w is the weight matrix, b is the bias term, \n  and y is the actual target value.\n  The function takes in the input data X, the actual target values y, the weight matrix w, and the bias term b as arguments. \n  It calculates the gradient of the log loss with respect to the weight matrix and returns the result.\n\n  Here is a breakdown of each part of the function:\n  X: This is the input data, which is an array of independent variables.\n  y: This is the actual target values, which is an array of the same size as X.\n  w: This is the weight matrix, which is an array of the same size as X.\n  b: This is the bias term, which is a scalar value.\n  gradient_dw: This is the gradient of the log loss with respect to the weight matrix. \n  It is calculated as  np.dot(X.T,sigmoid(np.dot(w.T, X )+ b)-y).\n  return gradient_dw: This returns the gradient of the log loss with respect to the weight matrix.\n\"\"\"\n  gradient_dw = np.dot(X.T,sigmoid(np.dot(w.T, X )+ b)-y)\n  return gradient_dw"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":true,"cellMetadata":{},"nuid":"9c3dd1ac-0aa2-4369-a644-63361dba5cff","inputWidgets":{},"title":"Gradient descent"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"code","source":["def prediction(X,w,b):\n  \"\"\"\n  This function takes in three arguments:\n  X, which is a list of values of the independent variables\n  w, which is the weight of the model\n  b, which is the bias of the model\n  It then iterates through each value in X, calculates the dot product of w and the current value of X, adds b, \n  and passes the result through the sigmoid function.\n If the output of the sigmoid function is greater than or equal to 0.5, the function appends 1 to the list predictions. \n Otherwise, it appends 0.\n\n Finally, the function returns the list predictions, which contains the predicted values of the dependent variable y.\n \"\"\"\n  predictions = []\n  for i in range(len(X)):\n    z = np.dot(w, X[i]) + b\n    y_pred = sigmoid(z)\n    if y_pred >= 0.5:\n      predictions.append(1)\n    else:\n      predictions.append(0)\n  return predictions   "],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":true,"cellMetadata":{},"nuid":"12664755-5983-4b52-814a-0b60d16df209","inputWidgets":{},"title":"Predicting 0's and 1's"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"code","source":["def logistic_regression(X_train,y_train,X_test,y_test,epochs,LR,L1,L2, batch_size = 1):\n    \n    ''' In this function, we will implement logistic regression with L1 and L2 regularization\n    X_train:Training Input Dataset\n    y_train:Training Output Dataset\n    X_test:Testing Input Dataset\n    y_test:Testing Output Dataset\n    epochs:No of epochs or iterations to fit the best weights\n    LR:Learning Rate for Gradient Descent\n    L1:L1 regularization hyperparameter\n    L2:L2 regularization hyperparameter\n    batch_size:Batch Size to implement SGD \n    '''\n    N = len(X_train)\n    train_loss = []\n    test_loss = []\n    w, b = initialize_weights(X_train[0])\n    for i in range(epochs):\n      \n      for i in range(0,N,batch_size):\n          dw = gradient_wrtw(X_train[i],y_train[i],w,b) + L1 * np.sign(w) + L2 * w  # add L1 and L2 regularization terms to gradient update for weights\n          db = gradient_wrtb(X_train[i],y_train[i],w,b)\n          w = w - (LR * dw)\n          b = b - (LR * db)\n      y_pred_train=prediction(X_train,w,b)\n      train_loss.append(logloss(y_test,y_pred_train))\n      y_pred_test=prediction(X_test,w,b)\n      test_loss.append(logloss(y_test,y_pred_test))\n    return w,b,train_loss,test_loss\n\n\n\n\n"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":false,"cellMetadata":{},"nuid":"94dec56d-2dd5-4ac6-9645-9a7021ed7e95","inputWidgets":{},"title":""}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"code","source":["epochs = 100\nLR = 0.001\nL1 = 0.01\nL2 = 0.01\n\n# Call the function\nw,b,train_loss,test_loss = logistic_regression(X_train,y_train,X_test,y_test,epochs,LR,L1,L2)"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":false,"cellMetadata":{},"nuid":"c74b18b6-6beb-44ea-9bba-529fb2cc72f8","inputWidgets":{},"title":""}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\">/local_disk0/tmp/1673174063994-0/PythonShell.py:27: RuntimeWarning: divide by zero encountered in log10\n  \n/local_disk0/tmp/1673174063994-0/PythonShell.py:27: RuntimeWarning: invalid value encountered in multiply\n  \n</div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\">/local_disk0/tmp/1673174063994-0/PythonShell.py:27: RuntimeWarning: divide by zero encountered in log10\n  \n/local_disk0/tmp/1673174063994-0/PythonShell.py:27: RuntimeWarning: invalid value encountered in multiply\n  \n</div>"]}}],"execution_count":0},{"cell_type":"code","source":["w"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":false,"cellMetadata":{},"nuid":"8f5c6a05-2f6d-4a0f-b0c5-0d9eb70d9e1c","inputWidgets":{},"title":""}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\">Out[59]: array([-0.38480577,  0.20085543, -0.01393337,  0.34277832, -0.17012719,\n        0.47125717, -0.35970298, -0.00918947,  0.10698503,  0.10306338,\n        0.09112999, -0.00086807, -0.00720535,  0.28264151, -0.03041991])</div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\">Out[59]: array([-0.38480577,  0.20085543, -0.01393337,  0.34277832, -0.17012719,\n        0.47125717, -0.35970298, -0.00918947,  0.10698503,  0.10306338,\n        0.09112999, -0.00086807, -0.00720535,  0.28264151, -0.03041991])</div>"]}}],"execution_count":0},{"cell_type":"code","source":["b"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":false,"cellMetadata":{},"nuid":"ac19afb5-e5b2-403f-a6cf-0cb29f24b2da","inputWidgets":{},"title":""}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\">Out[60]: -0.8424623769760174</div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\">Out[60]: -0.8424623769760174</div>"]}}],"execution_count":0},{"cell_type":"code","source":["y_pred_train = prediction(X_train,w,b)\ny_pred_test = prediction(X_test,w,b)"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":false,"cellMetadata":{},"nuid":"2e125f20-9b19-4237-96e5-f5627c0518d1","inputWidgets":{},"title":""}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"code","source":["cm_tr = confusion_matrix(y,y_pred)\ncm_tr"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":false,"cellMetadata":{},"nuid":"986195c8-b691-4dfc-bf51-ad6c0ee9bb6d","inputWidgets":{},"title":""}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\">Out[69]: array([[32279,  2625],\n       [ 5936,  9160]])</div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\">Out[69]: array([[32279,  2625],\n       [ 5936,  9160]])</div>"]}}],"execution_count":0},{"cell_type":"code","source":["plot_confusion_matrix(cm_tr)"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":false,"cellMetadata":{},"nuid":"d49084c0-79e6-4cdc-beaf-bb4629788d5e","inputWidgets":{},"title":""}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"/plots/6fb034d0-6cfe-4feb-922f-cb9c0201d534.png","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"image","arguments":{}}},"output_type":"display_data","data":{"image/png":"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"}}],"execution_count":0},{"cell_type":"code","source":["cm_te = confusion_matrix(y_test,y_pred_test)\ncm_te\n"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":false,"cellMetadata":{},"nuid":"37e1127c-f806-4242-9e8e-8d245de89adb","inputWidgets":{},"title":""}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\">Out[71]: array([[8082,  651],\n       [1475, 2292]])</div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\">Out[71]: array([[8082,  651],\n       [1475, 2292]])</div>"]}}],"execution_count":0},{"cell_type":"code","source":["plot_confusion_matrix(cm_te)"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":false,"cellMetadata":{},"nuid":"61248453-84d7-4508-8d1b-24f0d094c6bc","inputWidgets":{},"title":""}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"/plots/8fb5afd9-3a55-4b02-a53e-9780a4cb6d8d.png","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"image","arguments":{}}},"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAVEAAAEmCAYAAADbUaM7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nO3de7xWZZ338c93g3lWQIxBIHUm0qGDxjDiKcekEGyecDqoWUEOPdREdp5Jm3kiNZtqmix70iJlREsJT0llIlG+1B5PpGaKNTCaAqKIG/GAh8Df88e67rjd3fe9133aa+97f9+v13rtta51rbWutXnx29dhrWspIjAzs8Z0FV0AM7OBzEHUzKwJDqJmZk1wEDUza4KDqJlZExxEzcya4CBqZtYEB9E+JulkSSskPSNpvaSfSTqybP8ESUskbZb0tKRfSjq8bP9+kkLStT3O+31JX0jrR0t6KV2jtPw47btI0hd7HFs659C0faSk/5fK0C3pV5L+Nu37gKSbexz/AUm/lbRF0qOSzpc0rGz/F9L5TyhLG5rS9qvye7oh7T+oR/rVKf3oCmUISSem7TeV3fuzaV/57+NV6RrPp+2Nkq6SNLpHub+f1sdI2tTj32pcSptc6R5scHAQ7UOSPgV8A/gSMAp4FXAeMCPt/yvgV8Bvgf2BfYCrgeslHdbjdJPLg2sFj0TEbmXL/8pZxj2AnwDfAkYAY4AzgBeq5P808BXgn4E9gUOBfYFlkl5RlrUbOEPSkDzlSP4bmFl2rb2Aw4DHK+Sdla4xEyAibirdO/DalGdY2e/j4ZT20ZTn1cBuwNcqFSQi1gGfBS6QtFNK/i7wXxFxWx33ZB3GQbSPSNoTOBOYGxFXRcSzEfHHiPhxRPxzyvYF4JaI+NeI6I6IpyPiXOASskBV7qvA2W0o6msAIuKyiNgWEc9FxPURcU+Fe9qDLMCeGhHXpfv5A3ACsB/wvrLs1wEv9kjrzQ+AE8sC73vI/qi82KMc+wJ/B8wBjpX0F3VcA4CIeBL4EXBwjWzfA9YD8yTNAg4A/q3ea1lncRDtO4cBO5EFgWreClxeIX0xcISkncvSzgNeI+ktrSsikNX+tklaKGm6pOE18h5Odk9XlSdGxDPAtWT386dk4P+QBaAdcpblEWAlMDVtzwQurpBvJrAiIq4E7gfem/P8f5Jque8AVlfLE9k70h8EPkLWovjfEbGl3mtZZ3EQ7Tt7ARsjYmuNPCPJajo9rSf7txpRlvYcWU30ixXyA+wj6cmy5YQq+V4mIp4CjiQLet8DHk99tKOqlLfaPa1P+8vPvYSsKf7BPGVJLgZmSjqQrDl+S4U8M4FL0/qllHUB5HCupM3AxlTeU3vJ/xBZcH8KuLGO61iHchDtO08AI0uDN1VsBEZXSB8NvARs6pF+ATBKUqX+zkciYljZsjilbwV61gR3SOd/CSAi7o+ID0TEWOB1ZH2z36hS3mr3NDrt7+nfgH8lq8HmcRVwDPBRsm6Nl5F0BFn/8aKUdCnwekm1muXlPhYRewJvAIYDY3vJfxrZv+UG4DM5r2EdzEG079xCNjhzfI08PwfeXSH9BLK+0pc1HSPiRbI+ybMA5SzHw2T9leX2B9ZExEs9M0fE74CLyIJpT6V7ekd5oqTdgOnA8grnW0bWZP5InsKme/4Z8E9UCKJkA0oC7pb0KHBbWXpuEfFbslr9tyVV/F1KmkA2gPZBYDbwOUnj67mOdR4H0T4SEZuBz5P9Jz1e0i6Sdkj9jl9N2c4ADpd0tqQRknaXdCpZ8/SzVU59CVmtblrOolwJvE3SVElDJO1DVjtcBCDpQEmfljQ2bY8jG9C5tco9nQF8S9K0dD/7kfXhrqVy0IOsJvovOcsL8Dng79Kg1Z+kUfITyAaUDi5bTgVO7qXWX8lCsqcm3t5zh6Qu4ELgqxHxuzTQdi4wv1rQtcHBQbQPRcR/Ap8iC1qPA2vImqk/SvtXkfVHHgT8gaxf8Z3AsRHxqyrn3EYWnEdU2l8h/31kQfHfyR4JuoWs9nZGyvI0MBm4TdKzZMHzXuDTVc73VbIg9zWyfsLb0n1NiYiKj0Wle7k9T3lT/kci4uYKu44n6xu+OCIeLS3AAmAo+f+wlK7zIvBNsgGwnj4O7EL2VETJWcBfUF8fr3UYeVJmM7PGuSZqZtYEB1EzsyY4iJqZNcFB1MysCfU+AtJWO0mxu+N6x9j3jW8ougjWIn94+GE2bnyipY9yjdPQeJ78A9sbeWlpRFR94kLSJ8melAiySXxOIXvpYxHZG4O/Bt4fES9K2pHsbbi/IXt54sTSI3SSTid7Dngb2csYS2uVq18F0d3p4p3sUnQxrEW+c/MNRRfBWmTSkUe3/JzPE7yTXXPn/y5Pj6y2T9IY4GPAhIh4TtJi4CTgOOCciFgk6TtkwfH89HNTRLxa0klkE/ycmF6oOIls5q99gJ9Lek16lLAiV/vMrBAiC0B5lxyGAjunlyx2IXvO+hjgirR/IdvfGJyRtkn7p6SXJmYAiyLihYh4kOztukNqXdRB1MwK0yXlXsjmaVhRtswpnSfN9/o1stea1wObyZrvT5ZNkLOWbH5c0s816ditKf9e5ekVjqmoXzXnzWzwKNVE67AxIiZVPFc2ZeMMsnkgniSbUrKuN9Ya5SBqZoUZWs9QVe0xqLcAD0bE4wCSrgKOAIZJGppqm2OBdSn/OmAcsDY1//ckG2AqpZeUH1ORm/NmVgiRvynf1fscLw8Dh6aJfQRMIZvQ+5fAu1KeWcA1aX0J22f6ehfwizTp9hLgJEk7StofGE8v8zy4JmpmhWlVLS4ibpN0BXAn2Zy5dwHzgZ8Ci5R9nPEuspm4SD8vkbSabCKek9J57ksj+yvTeebWGpkHB1EzK4iArhY+eRoR84B5PZIfoMLoekQ8T+W5e4mIs6nj+2UOomZWmE7oT3QQNbNiCDphPmsHUTMrRAOPOPVLDqJmVphW9okWxUHUzArjmqiZWYMEDHWfqJlZY1r9iFNRHETNrDBuzpuZNaGLgV8VdRA1s0K4OW9m1iQ3583MGiS5Jmpm1hT3iZqZNcE1UTOzBvndeTOzJgj5jSUzs2a4OW9m1oQOiKEOomZWDD9sb2bWJD/iZGbWID9sb2bWpE54xKkT7sHMBijVsdQ8j3SApLvLlqckfULSCEnLJK1KP4en/JJ0rqTVku6RNLHsXLNS/lWSZvV2Dw6iZlaIbGBJuZdaIuL3EXFwRBwM/A2wBbgaOA1YHhHjgeVpG2A6MD4tc4DzASSNIPt2/WSy79XPKwXeahxEzawwXXUsdZgC/E9EPATMABam9IXA8Wl9BnBxZG4FhkkaDRwLLIuI7ojYBCwDptW6mPtEzawwdY4rjZS0omx7fkTMr5DvJOCytD4qItan9UeBUWl9DLCm7Ji1Ka1aelUOomZWGNX32ufGiJjUy/leAbwdOL3nvogISVFfCXvn5ryZFaKeQaU6Qu104M6IeCxtP5aa6aSfG1L6OmBc2XFjU1q19KocRM2sMG3oE30P25vyAEuA0gj7LOCasvSZaZT+UGBzavYvBaZKGp4GlKamtKrcnDezwrRyEidJuwJvBT5UlvxlYLGk2cBDwAkp/VrgOGA12Uj+KQAR0S3pLOCOlO/MiOiudV0HUTMrjFr42mdEPAvs1SPtCbLR+p55A5hb5TwLgAV5r+sgamaFqLOvs99yEDWzwjiImpk1wROQmJk1SMhT4ZmZNUytHZ0vioOomRWmA2Kog6iZFcfNeTOzBvkRJzOzJrlP1MysCR0QQx1Ezaw4rXztsygOomZWCH933sysSR0QQx1Ezaw4fsTJzKwJnTA637aZ7SUtkLRB0r3tuoaZDVyibV/77FPtLNtF9PKpUTMb3NrwjaU+17bmfETcKGm/dp3fzAa+Or/22S8V3icqaQ4wB2C3fv33xsxarRP+xxceRCNiPjAfYG8Nafk3oc2sf+rvzfS8Cg+iZjZISW7Om5k1oxPeWGrnI06XAbcAB0ham777bGYGpEechij30uv5pGGSrpD0O0n3SzpM0ghJyyStSj+Hp7ySdK6k1ZLukTSx7DyzUv5Vkmb1dt22BdGIeE9EjI6IHSJibERc2K5rmdkAlD4PknfJ4ZvAdRFxIHAQcD9wGrA8IsYDy9M2wHRgfFrmAOcDSBoBzAMmA4cA80qBt5r+/AyrmXU4pX7RPEsv59kTOAq4ECAiXoyIJ4EZwMKUbSFwfFqfAVwcmVuBYZJGA8cCyyKiOyI2Acvo5Xl3B1EzK0ydNdGRklaULXPKTrU/8DjwX5LuknSBpF2BURGxPuV5FBiV1scAa8qOX5vSqqVX5YElMytMnaPzGyNiUpV9Q4GJwKkRcZukb7K96Q5ARISklj9G6ZqomRVCtLRPdC2wNiJuS9tXkAXVx1IznfRzQ9q/DhhXdvzYlFYtvSoHUTMrhqBLyr3UEhGPAmskHZCSpgArgSVAaYR9FnBNWl8CzEyj9IcCm1OzfykwVdLwNKA0NaVV5ea8mRWmxc/anwr8QNIrgAeAU8gqiovTI5YPASekvNcCxwGrgS0pLxHRLeks4I6U78yI6K51UQdRMytIa99Yioi7gUp9plMq5A1gbpXzLAAW5L2ug6iZFUKAOqBD0UHUzIoh6OqA9z4dRM2sMJ6AxMysCR0QQx1EzawY2XfnB34UdRA1s2Lkn1ikX3MQNbPCuE/UzKwJHRBDHUTNrBild+cHOgdRMyuGhPycqJlZ44Y4iJqZNcbNeTOzJnl03sysUX5O1MysOa6Jmpk1oQNiqIOomRUjG1ga+FHUQdTMiiFPymxm1oTWfh6kKA6iZlYcP2xvZtYggboGfnveQdTMitMBNdGB/2fAzAao9LR93qW3s0l/kPRbSXdLWpHSRkhaJmlV+jk8pUvSuZJWS7pH0sSy88xK+VdJmtXbdavWRCV9C4hq+yPiY73elZlZFRLtmMXpzRGxsWz7NGB5RHxZ0mlp+7PAdGB8WiYD5wOTJY0A5pF9vz6AX0taEhGbql2wVnN+RVO3YmbWm/aPzs8Ajk7rC4EbyILoDODiiAjgVknDJI1OeZdFRHdWPC0DpgGXVbtA1SAaEQvLtyXtEhFbGr0TM7OeWlwTDeB6SQF8NyLmA6MiYn3a/ygwKq2PAdaUHbs2pVVLr6rXPlFJh0laCfwubR8k6bze78fMrBf19YmOlLSibJnT42xHRsREsqb6XElHle9Mtc6qXZSNyjM6/w3gWGBJKshvehbOzKxuUr2j8xsjYlK1nRGxLv3cIOlq4BDgMUmjI2J9aq5vSNnXAePKDh+b0taxvflfSr+hVqFyjc5HxJoeSdvyHGdmVouk3Esv59lV0u6ldWAqcC9Z5a80wj4LuCatLwFmplH6Q4HNqdm/FJgqaXgayZ+a0qrKUxNdI+lwICTtAHwcuD/HcWZmtbWuT3QUcHUKtkOBSyPiOkl3AIslzQYeAk5I+a8FjgNWA1uAUwAiolvSWcAdKd+ZpUGmavIE0Q8D3yTrXH2ELCrPzX9vZmYVCDSkNUE0Ih4ADqqQ/gQwpUJ6UCWORcQCYEHea/caRNMzV+/Ne0Izs9w6YAKSPKPzfynpx5Iel7RB0jWS/rIvCmdmHSx9Mjnv0l/lGVi6FFgMjAb2AS6nxoOnZma5tfC1z6LkCaK7RMQlEbE1Ld8Hdmp3wcxsEOhS/qWfqvXu/Ii0+rP0zukisgdVTyQb2TIza1hWwey/wTGvWgNLvyYLmqW7/FDZvgBOb1ehzGyQ6Mc1zLxqvTu/f18WxMwGm/7d15lXrkmZJb0OmEBZX2hEXNyuQpnZ4NDpzXkAJM0je5d0Allf6HTgZsBB1MwaJ9CQgT8vfJ47eBfZE/+PRsQpZG8F7NnWUpnZ4NDJo/NlnouIlyRtlbQH2Swo43o7yMyspn7+/GdeeYLoCknDgO+Rjdg/A9zS1lKZ2aDQn99EyivPu/MfSavfkXQdsEdE3NPeYpnZoNDJNdHyr99V2hcRd7anSGY2KIh+3deZV62a6H/W2BfAMS0uC/u+/q85/6eLW31aK8i2O39edBGsVbY81ZbTdvQjThHx5r4siJkNNv171D2vXA/bm5m1RSfXRM3M2ko4iJqZNU4wZEjRhWhanpntJel9kj6ftl8l6ZD2F83MOt4gmZT5POAw4D1p+2ng220rkZkNDqXm/AAPonma85MjYqKkuwAiYpOkV7S5XGY2GPTj4JhXniD6R0lDyJ4NRdLewEttLZWZDQKCrsExi9O5wNXAKyWdTTYN3pfaWiozGxxa3JyXNETSXZJ+krb3l3SbpNWSflhqRUvaMW2vTvv3KzvH6Sn995KO7e2avQbRiPgB8C/AvwPrgeMj4vJcd2RmVk17+kQ/Dtxftv0V4JyIeDWwCZid0mcDm1L6OSkfkiYAJwGvBaYB56WWeFV5RudfBWwBfgwsAZ5NaWZmzWlhEJU0FngbcEHaFtnr6VekLAuB49P6jLRN2j8l5Z8BLIqIFyLiQWA1UPNppDx9oj9l+wfrdgL2B35PFqnNzBpUd5/oSEkryrbnR8T8su1vkLWad0/bewFPRsTWtL0WGJPWxwBrACJiq6TNKf8Y4Nayc5YfU1GeqfBeX76dZnf6SJXsZmb5iHqD6MaImFTxVNLfAxsi4teSjm5B6XKr+42liLhT0uR2FMbMBpnWPeJ0BPB2SceRtZj3AL4JDJM0NNVGxwLrUv51ZF/oWCtpKNknj54oSy8pP6aiPB+q+1TZZhcwEXgkx02ZmVUlhFr0iFNEnA6cDpBqop+JiPdKupzsO3GLgFnANemQJWn7lrT/FxERkpYAl0r6OrAPMB64vda189REdy9b30rWR3plvlszM6uh/Q/bfxZYJOmLwF3AhSn9QuASSauBbrIReSLiPkmLgZVk8W5uRGyrdYGaQTQN7e8eEZ9p6jbMzHpq0yxOEXEDcENaf4AKo+sR8Tzw7irHnw2cnfd6tT4PMjSNWh2R92RmZnXp8Nc+byfr/7w79RNcDjxb2hkRV7W5bGbW0Trjtc88faI7kY1aHcP250UDcBA1s+Z0eE30lWlk/l62B8+SaGupzKzzDYKZ7YcAu/Hy4FniIGpmzevwILo+Is7ss5KY2SDTGZ8HqRVEB/6fCDPrvwZBc35Kn5XCzAahDh+dj4juviyImQ1CHV4TNTNrLwdRM7MGDYI+UTOzNurwPlEzs7ZzTdTMrAkOomZmDRIgN+fNzBrU+W8smZm1l5vzZmYNkkfnzcya45qomVkTPLBkZtYE10TNzBrUIX2iA/8OzGzgkvIvNU+jnSTdLuk3ku6TdEZK31/SbZJWS/qhpFek9B3T9uq0f7+yc52e0n8v6djebsFB1MyKo678S20vAMdExEHAwcA0SYcCXwHOiYhXA5uA2Sn/bGBTSj8n5UPSBOAk4LXANOA8STUfZnUQNbNiKD1sn3epITLPpM0d0hJkXym+IqUvBI5P6zPSNmn/FElK6Ysi4oWIeBBYDRxS69oOomZWnPqa8yMlrShb5rz8VBoi6W5gA7AM+B/gyYjYmrKsBcak9THAGoC0fzOwV3l6hWMq8sCSmRVE9T7itDEiJlXbGRHbgIMlDQOuBg5ssoC5uCZqZsUQ0KX8S04R8STwS+AwYJikUmVxLLAura8DxgGk/XsCT5SnVzimIgdRMytO60bn9041UCTtDLwVuJ8smL4rZZsFXJPWl6Rt0v5fRESk9JPS6P3+wHjg9lrXdnPezIrTujeWRgML00h6F7A4In4iaSWwSNIXgbuAC1P+C4FLJK0GuslG5ImI+yQtBlYCW4G5qZugKgdRMyuG6mum1xIR9wBvrJD+ABVG1yPieeDdVc51NnB23ms7iJpZcfzuvJlZE/zuvJlZo+p+xKlfchA1s2IIfx7EzKwpbs6bmTXKU+H1StK0NJ3UakmntfNaZjbAiJY9bF+kttVE00Ov3yZ7c2AtcIekJRGxsl3XNLMBpgMGltp5B4cAqyPigYh4EVhENs2UmRnZ6PzAr4m2M4jmmlJK0pzS1FaPd3e3sThm1u90deVf+qnCSxYR8yNiUkRM2nvEiKKLY2Z9xX2ivap7SikzG0z8sH1v7gDGp+mk1pHNknJyG69nZgNNlx+2ryoitkr6KLAUGAIsiIj72nU9MxtgWjiLU5Ha+rB9RFwLXNvOa5jZAObmvJlZE/rxgFFeDqJmVhAPLJmZNUWuiZqZNUi4Jmpm1jg3583MmuNHnMzMmtABNdGBfwdmNjBJ2RtLeZeap9I4Sb+UtFLSfZI+ntJHSFomaVX6OTylS9K5aa7jeyRNLDvXrJR/laRZvd2Gg6iZFad1E5BsBT4dEROAQ4G5kiYApwHLI2I8sDxtA0wHxqdlDnB+VhyNAOYBk8mm85xXCrzVOIiaWXHUlX+pISLWR8Sdaf1p4H6yqTdnAAtTtoXA8Wl9BnBxZG4FhkkaDRwLLIuI7ojYBCwDptW6tvtEzawY9b87P1LSirLt+REx/89Pq/2ANwK3AaMiYn3a9SgwKq1Xm+841zzI5RxEzaw49Q0sbYyISTVPJ+0GXAl8IiKeKn+YPyJCUjRUzhrcnDez4rRwUmZJO5AF0B9ExFUp+bHUTCf93JDSq813XPc8yA6iZlYQtaxPVFmV80Lg/oj4etmuJUBphH0WcE1Z+sw0Sn8osDk1+5cCUyUNTwNKU1NaVW7Om1lxWvfu/BHA+4HfSro7pX0O+DKwWNJs4CHghLTvWuA4YDWwBTgFICK6JZ1FNqk8wJkRUfPjbw6iZlaMFr47HxE3pzNWMqVC/gDmVjnXAmBB3ms7iJpZQYSG+PMgZmaN64DXPh1EzawYpU8mD3AOomZWEE+FZ2bWHNdEzcya0OWaqJlZY3K+idTfOYiaWXHcJ2pm1gTXRM3MmuEgambWIHlgycysOa6Jmpk1xm8smZk1aeDHUAdRMyvSwI+iDqJmVhA/bG9m1hwHUTOzZjiImpk1zjVRM7NmOIiamTVGfmPJzKxJA78mOvD/DJjZgCUp95LjXAskbZB0b1naCEnLJK1KP4endEk6V9JqSfdImlh2zKyUf5WkWb1d10HUzIpTmpg5z9K7i4BpPdJOA5ZHxHhgedoGmA6MT8sc4PysOBoBzAMmA4cA80qBtxoHUTMriOpcaouIG4HuHskzgIVpfSFwfFn6xZG5FRgmaTRwLLAsIrojYhOwjD8PzC/jPlEzK059jziNlLSibHt+RMzv5ZhREbE+rT8KjErrY4A1ZfnWprRq6VU5iJpZMeqfxWljRExq9HIREZKi0eOrcXPezArUuuZ8FY+lZjrp54aUvg4YV5ZvbEqrll6Vg6iZFae1A0uVLAFKI+yzgGvK0memUfpDgc2p2b8UmCppeBpQmprSqnJz3swK0tpZnCRdBhxN1ne6lmyU/cvAYkmzgYeAE1L2a4HjgNXAFuAUgIjolnQWcEfKd2ZE9BysehkHUTMrTguDaES8p8quKRXyBjC3ynkWAAvyXtdB1MyK4c+DmJk1y0HUzKxxromamTXKnwcxM2vSwA+iygap+gdJj5M9htDpRgIbiy6EtcRg+bfcNyL2buUJJV1H9vvLa2NE1HyPvQj9KogOFpJWNPP6mvUf/rc0v7FkZtYEB1EzsyY4iBajt+m7bODwv+Ug5z5RM7MmuCZqZtYEB1EzsyY4iJqZNcFBtA9JGlJ0Gax5kg6QdJikHfxvah5Y6gOSXhMR/53Wh0TEtqLLZI2R9A7gS2z/lMQK4KKIeKrQgllhXBNtM0l/D9wt6VKAiNjm2svAJGkH4ERgdkRMIfvUxDjgs5L2KLRwVhgH0TaStCvwUeATwIuSvg8OpAPcHsD4tH418BNgB+BkqQOmJLK6OYi2UUQ8C/wjcCnwGWCn8kBaZNmsfhHxR+DrwDskvSkiXgJuBu4Gjiy0cFYYB9E2i4hHIuKZiNgIfAjYuRRIJU2UdGCxJbQ63QRcD7xf0lERsS0iLgX2AQ4qtmhWBM8n2oci4glJHwL+Q9LvgCHAmwsultUhIp6X9AMggNPTH8EXgFHA+kILZ4VwEO1jEbFR0j3AdOCtEbG26DJZfSJik6TvASvJWhfPA++LiMeKLZkVwY849TFJw4HFwKcj4p6iy2PNSQOEkfpHbRByEC2ApJ0i4vmiy2FmzXMQNTNrgkfnzcya4CBqZtYEB1EzsyY4iJqZNcFBtENI2ibpbkn3Srpc0i5NnOsiSe9K6xdImlAj79GSDm/gGn+Q9GffHK+W3iPPM3Ve6wuSPlNvGc3ycBDtHM9FxMER8TrgReDD5TslNfRiRUR8MCJW1shyNFB3EDXrFA6inekm4NWplniTpCXASklDJP2HpDsk3ZNeQUWZ/yvp95J+DryydCJJN0ialNanSbpT0m8kLZe0H1mw/mSqBb9J0t6SrkzXuEPSEenYvSRdL+k+SRcAvc54JOlHkn6djpnTY985KX25pL1T2l9Jui4dc5PnJbC+4Nc+O0yqcU4HrktJE4HXRcSDKRBtjoi/lbQj8CtJ1wNvBA4AJpC9A74SWNDjvHsD3wOOSucaERHdkr4DPBMRX0v5LgXOiYibJb0KWAr8NTAPuDkizpT0NmB2jtv5x3SNnYE7JF0ZEU8AuwIrIuKTkj6fzv1Rss8XfzgiVkmaDJwHHNPAr9EsNwfRzrGzpLvT+k3AhWTN7Nsj4sGUPhV4Q6m/E9iTbG7Mo4DL0vR8j0j6RYXzHwrcWDpXRHRXKcdbgAllU2vuIWm3dI13pGN/KmlTjnv6mKR/SOvjUlmfAF4CfpjSvw9cla5xOHB52bV3zHENs6Y4iHaO5yLi4PKEFEyeLU8CTo2IpT3yHdfCcnQBh/Z8rbXe+YolHU0WkA+LiC2SbgB2qpI90nWf7Pk7MGs394kOLkuBf0qfuUDSa9Ls+zcCJ6Y+09FUnp7vVuAoSfunY0ek9KeB3cvyXQ+cWtqQVApqNwInp7TpwPBeyronsCkF0APJasIlXUCpNn0yWTfBU8CDkt6driFJnt/T2s5BdHC5gKy/805J9wLfJWuNXA2sSvsuBm7peWBEPA7MIWs6/4btzekfA/9QGlgCPgZMSgNXK9n+lMAZZEH4PrJm/cO9lPU6YKik+4EvkwXxkmeBQ9I9HH0UzuwAAABESURBVAOcmdLfC8xO5bsPmJHjd2LWFE9AYmbWBNdEzcya4CBqZtYEB1EzsyY4iJqZNcFB1MysCQ6iZmZNcBA1M2vC/wegOB6WKDnt0wAAAABJRU5ErkJggg=="}}],"execution_count":0},{"cell_type":"code","source":["cr_tr = classification_report(y_train,y_pred_train)\nprint(cr_tr)"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":false,"cellMetadata":{},"nuid":"63f33124-e6e3-48e7-925c-bbef11aae626","inputWidgets":{},"title":""}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\">              precision    recall  f1-score   support\n\n           0       0.84      0.92      0.88     26171\n           1       0.78      0.61      0.68     11329\n\n    accuracy                           0.83     37500\n   macro avg       0.81      0.77      0.78     37500\nweighted avg       0.82      0.83      0.82     37500\n\n</div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\">              precision    recall  f1-score   support\n\n           0       0.84      0.92      0.88     26171\n           1       0.78      0.61      0.68     11329\n\n    accuracy                           0.83     37500\n   macro avg       0.81      0.77      0.78     37500\nweighted avg       0.82      0.83      0.82     37500\n\n</div>"]}}],"execution_count":0},{"cell_type":"code","source":["cr_te = classification_report(y_test,y_pred_test)\nprint(cr_te)"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":false,"cellMetadata":{},"nuid":"3da1a0e5-8a1b-4616-b7e7-2265e7f88a34","inputWidgets":{},"title":""}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\">              precision    recall  f1-score   support\n\n           0       0.85      0.93      0.88      8733\n           1       0.78      0.61      0.68      3767\n\n    accuracy                           0.83     12500\n   macro avg       0.81      0.77      0.78     12500\nweighted avg       0.83      0.83      0.82     12500\n\n</div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\">              precision    recall  f1-score   support\n\n           0       0.85      0.93      0.88      8733\n           1       0.78      0.61      0.68      3767\n\n    accuracy                           0.83     12500\n   macro avg       0.81      0.77      0.78     12500\nweighted avg       0.83      0.83      0.82     12500\n\n</div>"]}}],"execution_count":0},{"cell_type":"code","source":["import matplotlib.pyplot as plt\nfrom sklearn.metrics import roc_auc_score, roc_curve\n\n# Compute ROC AUC score\nroc_auc = roc_auc_score(y_test, y_pred_test)\n\n# Compute false positive rate and true positive rate\nfpr, tpr, thresholds = roc_curve(y_test, y_pred_test)\n\n# Plot ROC curve\nplt.plot(fpr, tpr, label='ROC curve (area = %0.2f)' % roc_auc)\nplt.plot([0, 1], [0, 1], 'k--')  # random predictions curve\nplt.xlim([0.0, 1.0])\nplt.ylim([0.0, 1.0])\nplt.xlabel('False Positive Rate or (1 - Specificity)')\nplt.ylabel('True Positive Rate or (Sensitivity)')\nplt.title('Receiver Operating Characteristic')\nplt.legend(loc=\"lower right\")\nplt.show()"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":false,"cellMetadata":{},"nuid":"c6325c14-1dfa-4e73-b5d7-ef4c91094587","inputWidgets":{},"title":""}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"/plots/d414f93a-7a97-463c-8a1e-036a157caff1.png","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"image","arguments":{}}},"output_type":"display_data","data":{"image/png":"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"}}],"execution_count":0},{"cell_type":"code","source":["#end of notebook"],"metadata":{"application/vnd.databricks.v1+cell":{"showTitle":false,"cellMetadata":{},"nuid":"c7758316-8d76-44bb-a8c1-18565b8f76a7","inputWidgets":{},"title":""}},"outputs":[],"execution_count":0}],"metadata":{"application/vnd.databricks.v1+notebook":{"notebookName":"Logistic_model_creation","dashboards":[],"notebookMetadata":{"pythonIndentUnit":2},"language":"python","widgets":{},"notebookOrigID":2561119894750365}},"nbformat":4,"nbformat_minor":0}