Ch. 1 - Introduction to Advanced Dimensionality Reduction

Exploring the MNIST dataset

Exploring MNIST dataset

Digits features

Distance metrics

Euclidean distance

Minkowsky distance

KL divergence

PCA and t-SNE

Generating PCA from MNIST sample

t-SNE output from MNIST sample

Ch. 2 - Introduction to t-SNE

Building a t-SNE embedding

Computing t-SNE

Understanding t-SNE output

Optimal number of t-SNE iterations

Reproducing results

Optimal number of iterations

Effect of perplexity parameter

Perplexity of MNIST sample

Perplexity of bigger MNIST dataset

Classifying digits with t-SNE

Plotting spatial distribution of true classes

Computing the centroids of each class

Computing similarities of digits 1 and 0

Plotting similarities of digits 1 and 0

Ch. 3 - Using t-SNE with Predictive Models

Credit card fraud detection

Exploring credit card fraud dataset

Generating training and test sets

Training random forests models

Training a random forest with original features

Computing and visualising the t-SNE embedding

Training a random forest with embedding features

Predicting data

Predicting data using original features

Predicting data using embedding random forest

Visualizing neural networks layers

Exploring neural network layer output

Using t-SNE to visualise a neural network layer

Ch. 4 - Generalized Low Rank Models (GLRM)

Exploring fashion MNIST dataset

Exploring fashion MNIST

Visualizing fashion MNIST

Generalized Low Rank Models (GLRM)

Reducing data with GLRM

Improving model convergence

Visualizing a GLRM model

Visualizing the output of GLRM

Visualizing the prototypes

Dealing with missing data and speeding-up models

Imputing missing data

Training a random forest with original data

Training a random forest with compressed data

Summary of the course

About Michael Mallari

Michael is a hybrid thinker and doer—a byproduct of being a StrengthsFinder “Learner” over time. With nearly 20 years of engineering, design, and product experience, he helps organizations identify market needs, mobilize internal and external resources, and deliver delightful digital customer experiences that align with business goals. He has been entrusted with problem-solving for brands—ranging from Fortune 500 companies to early-stage startups to not-for-profit organizations.

Michael earned his BS in Computer Science from New York Institute of Technology and his MBA from the University of Maryland, College Park. He is also a candidate to receive his MS in Applied Analytics from Columbia University.

LinkedIn | Twitter | michaelmallari.com