Week-wise Modules

• Module 1: Understanding the practicality of data preprocessing: Hands-On - Import Library to Read and Slice the data, Understand the dataset and features mapping, Hands-On - Handling Missing Values, Label-Encoding for Categorical Data, Hands-On Label Encoding, On- Hot-Encoding for Categorical Data, Data normalization, Data Normalization using Standard Scaler, Data Normalization using min max, Train and Test Data Split explained.
• Module 2: Regression model: Simple Linear Regression, Ordinary Least Square and Regression Errors, Train and Test Model, Test the model and Predict Y Values, R-Squared and its Importance, Score and Get coefficients, Calculate RMSE (Root Mean Squared Error), Multiple Linear Regression, Degrees of Freedom, Adjusted R-Squared, Assumptions of Multiple Linear Regression, Linearity and Multicollinearity Assumption, Autocorrelation, Dummy Variable trap, Calculate and measure RMSLE for Kaggle.
• Module 3: Equipping the basis of classification: Logistic Regression, a Case study of the Kaggle Loan Prediction System, Confusion, Matrix, Mathematical Intuition of SVM, SVM Kernel Functions, SVM Kernel Function Types, IRIS Classification Problem. Multiple Model Creation and Comparison, Train and creating Models, Intuition Behind Decision Trees, Adult Income Prediction Problem Understanding (Case study implementation), Split data and Import Classifier.
• Module 4: Let’s evaluate the classification model: Need for Evaluation and Accuracy Paradox, Classification Evaluation Measures Evaluation Metrics for Loan Prediction projects, Threshold and Adjusting Thresholds, AUC ROC Curve using Python, Univariate Feature Selection, F-Test for Regression and Classification, Chi-Squared Intuition, Dimensionality Reduction, Principal Components Analysis Convert the Dimensions to PCA, Regularization, L1 and L2 for Multicollinearity and Feature Selection.
• Module 5: Understanding the concept of cross-validation and hyperparameter: Cross Validation, Hyperparameter Tuning, Grid Search and Randomized Search Approach, GridSearchCV Parameters, Create GirdSearchCV Object, Select Best Model, Randomized Search Keras and TensorFlow, Activation Functions, Softmax Activation Function, Loss Function, Cross Entropy Loss Functions, Gradient Descent, Stochastic Gradient Descent, SGD with Momentum, Adagrad and RMSProp for Learning rate decay, Adam Optimizers Keras Neural Network Mode, Transfer learning, CNN, Optimization-based algorithms for Deep Neural networks.