background

Machine Learning Engineering

Machine Learning Engineering
1. Supervised Learning :-
Regression

Linear Regression
Logistic Regression
Polynomial Regression
Ridge Regression & Lasso Regression
Working
Math behind the Intuition
Learning the concepts of Coefficient and Residuals
Cost function
Feature scaling
Non-Linearity and non-Invertibility
Optimizing Linear Functions
Standard Error
Gradient decent intuation
Hypothesis Representation
Regularized Regressions
Regularization
L1 and L2 Regularization
Filter method
Wrapper method
Embedded Method
Decision Boundary
Case study using SciKit Learn

Classification
KNN

Intuition
Eager and Lazy Classifiers
Other names of KNN classifiers
How to Choose k?
Distance metrics used in KNN
Mathematically Demystifying KNN Algorithm
Weighted KNN
Characteristics of KNN Algorithm
Strength and weakness
Weighted KNN
Improvements of KNN performance
Fuzzy KNN
Case Study using SciKit Learn
Applying cross validation techniques and analyzing the Algorithm behaviour.
Improvisation on the Algorithm


Intuition
Visualize in Vector space
Large Margin Intuition
Significance of Binary Labels [+1,-1]
Inequalities and region
Maximum Margin: Formalization
Linear Support Vector machine
Non-Linear SVM
Hard Margin and Soft Margin
Kernel Tricks
C parameter?
Decision Functions
Multiclass Problem
Challenges on Multiclass classification
Polynomial Kernel
Gaussian RBF Kernel
SVR
Kernelized SVM
Tweak Performance
Upweighting
Drift Problem
Case Study using SciKit Learn
Strength and weakness


Intuition
Demystifying Probability
Conditional Probability
Bayes Theorem
Estimation of probability for the Dataset
Likelihoods
Gaussian, Bernoulli, Multinomial.
Discriminant Functions
Expectation Maximization Algorithm –EM
Case Study using SciKit Learn
Strength and weakness


Intuition
Training and Visualization
Predictions
Estimating Class Probabilities
Computational Complexity
CART Algorithm
HUNTS Algorithm
Gini Index, Entropy and Classification Error
Bagging and Bootstrapping
Regularization Hyperparameters
Case Study using SciKit Learn
Data Fragmentation
Tree Replication


Intuition
Voting Classifiers
Bagging and Pasting in Scikit-Learn
Out-of-Bag Evaluation
Random Patches and Random Subspaces
Random Forests
Boosting
AdaBoost
Gradient Boosting
Stacking
XGBoost
Feature Importance
Advantages and Disadvantages of The Algorithm
Performance Evaluation
Case Studies using Scikit-Learn

2. Unsupervised Learning :-

Introduction to clustering
Types of Clustering
Optimizing Objective
Data Characteristics
Prototype Based Approach
o K Means
Improvised K-Means Paper Implementations
Graph Based Approach
o Hierarchical Clustering
Density Based Approach
o DBSCAN


Intuition
Prototype Based Approach
Mathematically Demystifying KNN Algorithm
Expoloring K
Elbow Method
Characteristics of K-Means Clustering
Random Initialization
Data compression techniques
Distance Metrics for K Means
Strength and Weakness
Time and Space Complexity
Performance Evaluation
Improvised K-Means Implementations
Case Study using SciKit-Learn


Intuition
Graph based approach
Agglomerative and Divisive
Dendrograms
Proximity Methods
Strength and Weakness
Time and Space Complexity


Intuition
Density Based Approach
Mathematically Demystifying DBSCAN Algorithm
Analyzing Core points Border Point and Noise Points
Clustering Tendency
Cluster Evaluation Metrics
Cohesion and Separation
Silhouette coefficient
Time and space complexity
Strength and Weakness
Case Study using SciKit-Learn


Euclidian Distances
Squared Euclidian Distances
Manhattan Distance
Minkowski distance
Cosine measure
Jaccard distance


Association Analysis –Apriori Algorithm
Anomaly Detection