Introduction to machine learning, How machine learning is different
from general programming, Types of machine learning: supervised,
unsupervised, reinforcement, semi-supervised, Motivation for
machine learning, Applications of machine learning, Roadmap of a
machine learning project, What is Exploratory data analysis (EDA)
and its Goal, Activities under EDA, Data preprocessing, Data
preprocessing tasks: Data cleaning, Data transformation, Data
reduction, Data splitting, Handling imbalanced data, Feature
selection techniques, Loss functions: Mean square error, Mean
absolute error, Hinge loss, Binary cross entropy, Categorical cross
entropy, Optimization algorithm: Gradient descent algorithm, Types
of gradient descent: Batch, Stochastic, Mini batch, Mini batch with
momentum, Performance parameters: R2 score, Accuracy,
Precision, Recall, False positive rate, F-score, Receiver operating
characteristic (ROC) curve, Area Under the Curve (AUC)
What is regression, Types: Simple linear regression, Multiple linear
regression, Polynomial regression, Mathematical proof of linear
regression, Least squares estimation, Least squares regression-Line
of best fit, Illustration, Coefficient of determination (R-squared),
Adjusted R-squared, Bias, Variance, Overfitting, Underfitting,
Regularization: L1/Lasso, L2/Ridge, Elastic Net, Logistic regression
(Binary and Multiclass).
Classification and general approach, Classification algorithms, K-
Nearest neighbors: Euclidian and Manhattan distance, Bayes
theorem, Naïve bayes algorithm: Multinomial and Gaussian Naïve
bayes, Support Vector Machine: Support vectors, Hyper plane,
Marginal distance, Kernel, Soft & Hard margin, Neural networks,
Activation functions, Feedforward neural network, Multi-layer
perceptron, Backprop algorithm, Recurrent or feedback architecture,
Perceptron rule, Multilayer networks and back propagation
algorithm
Decision trees, Attribute selection measure: Entropy, Information
gain, Ginni impurity, ID3 algorithm, Bagging: Random Forest,
boosting: Adaptive boosting (Adaboost), Gradient boosting,
Extreme gradient boosting (XG Boost)
Clustering, Clustering algorithms, more common clustering
situation, General applications of clustering
Statistics associated with cluster analysis: Centroid, Dissimilarity
measure, Linkage methods,
Clustering as a pre-processing tool: Dimensionality Reduction and
Outlier detection, Principal component analysis (PCA)
Hard vs. soft clustering, Similarity, and dissimilarity between
objects
Type of data in clustering analysis: Binary variables, Nominal
variables, Ordinal variables,
Major clustering approaches: Partitioning method, Hierarchical
method, Density based method,
Types of clusters: Exclusive, Overlapping, Hierarchical,
Cluster centroid and distances,
Hierarchical clustering: Time and Space requirements, Hierarchical
Agglomerative Clustering (HAC): Linkage method, Variance and
Centroid method,
Cluster distance measures: Single link agglomerative clustering,
Complete-link clustering, Average-link clustering
K - means clustering, Importance of choosing initial centroids,
The K-medoids clustering method: PAM (Partitioning Around
Medoids), CLARA (Clustering Large Applications), CLARANS
(Randomized CLARA),
Density based clustering methods, DBSCAN: Density Based Spatial
Clustering of Applications with Noise, When DBSCAN Does NOT
Work Well,
External criteria for clustering quality, Different aspects of cluster
validation, Measures of cluster validity, measuring cluster validity
via correlation, Using similarity matrix for cluster validation,
Internal measures: SSE, Framework for cluster validity, Cohesion
and Separation, Silhouette coefficient.
