Machine Learning Course Details

1. Machine Learning

⇒ Introduction to Machine Learning & Predictive Modeling

⇒ Types of Business problems - Mapping of Techniques - Regression vs. classification vs. segmentation vs. Forecasting

⇒ Major Classes of Learning Algorithms -Supervised vs Unsupervised Learning

⇒ Different Phases of Predictive Modeling (Data Pre-processing, Sampling, Model Building, Validation)

⇒ Overfitting (Bias-Variance Trade off) & Performance Metrics

⇒ Feature engineering & dimension reduction

⇒ Concept of optimization & cost function

⇒ Overview of gradient descent algorithm

⇒ Overview of Cross validation(Bootstrapping, K-Fold validation etc)

⇒ Model performance metrics (R-square, Adjusted R-squre, RMSE, MAPE, AUC, ROC curve, recall, precision, sensitivity, specificity, confusion metrics )

2. Unsupervised Learning: Segmentation

⇒ What is segmentation & Role of ML in Segmentation?

⇒ Concept of Distance and related math background

⇒ K-Means Clustering

⇒ Expectation Maximization

⇒ Hierarchical Clustering

⇒ Spectral Clustering (DBSCAN)

⇒ Principle component Analysis (PCA)

3. Decision Tree

⇒ Decision Trees - Introduction - Applications

⇒ Types of Decision Tree Algorithms

⇒ Construction of Decision Trees through Simplified Examples; Choosing the "Best" attribute at each Non-Leaf node; Entropy; Information Gain, Gini Index, Chi Square, Regression Trees

⇒ Generalizing Decision Trees; Information Content and Gain Ratio; Dealing with Numerical Variables; other Measures of Randomness

⇒ Pruning a Decision Tree; Cost as a consideration; Unwrapping Trees as Rules

⇒ Decision Trees - Validation

⇒ Overfitting - Best Practices to avoid

4. Ensemble Learning (Supervised)

⇒ Concept of Ensembling

⇒ Manual Ensembling Vs. Automated Ensembling

⇒ Methods of Ensembling (Stacking, Mixture of Experts)

⇒ Bagging (Logic, Practical Applications)

⇒ Random forest (Logic, Practical Applications)

⇒ Boosting (Logic, Practical Applications)

⇒ Ada Boost

⇒ Gradient Boosting Machines (GBM)

⇒XGBoost

5. Artificial Neural Networks

⇒ Motivation for Neural Networks and Its Applications

⇒ Perceptron and Single Layer Neural Network, and Hand Calculations

⇒ Learning In a Multi Layered Neural Net: Back Propagation and Conjugant Gradient Techniques

⇒ Neural Networks for Regression

⇒ Neural Networks for Classification

⇒ Interpretation of Outputs and Fine tune the models with hyper parameters

⇒ Validating ANN models

6. Support Vector Machines

⇒ Motivation for Support Vector Machine & Applications

⇒ Support Vector Regression

⇒ Support vector classifier (Linear & Non-Linear)

⇒ Mathematical Intuition (Kernel Methods Revisited, Quadratic Optimization and Soft Constraints)

⇒ Interpretation of Outputs and Fine tune the models with hyper parameters

⇒ Validating SVM models

7. K-Nearest Neighbors Algorithm (KNN)

⇒ What is KNN & Applications?

⇒ KNN for missing treatment

⇒ KNN For solving regression problems

⇒ KNN for solving classification problems

⇒ Validating KNN model

⇒ Model fine tuning with hyper parameters

8. Naïve Bayes

⇒ Concept of Conditional Probability

⇒ Bayes Theorem and Its Applications

⇒ Naïve Bayes for classification

⇒ Applications of Naïve Bayes in Classifications

9. Data Mining

⇒ Taming big text, Unstructured vs. Semi-structured Data; Fundamentals of information retrieval, Properties of words; Creating Term-Document (TxD);Matrices; Similarity measures, Low-level processes (Sentence Splitting; Tokenization; Part-of-Speech Tagging; Stemming; Chunking)

⇒ Finding patterns in text: text mining, text as a graph

⇒ Natural Language processing (NLP)

⇒ Text Analytics – Sentiment Analysis using Python

⇒ Text Analytics – Word cloud analysis using Python

⇒ Text Analytics - Segmentation using K-Means/Hierarchical Clustering

⇒ Text Analytics - Classification (Spam/Not spam)

⇒ Applications of Social Media Analytics

⇒ Metrics(Measures Actions) in social media analytics

⇒ Examples & Actionable Insights using Social Media Analytics

⇒ Important python modules for Machine Learning (SciKit Learn, stats models, scipy, nltk etc)

⇒ Fine tuning the models using Hyper parameters, grid search, piping etc.

10. Project work

⇒ Applying different algorithms to solve the business problems and bench mark the results