Intro. of Dimensionality Reduction

TianHao Liu
Ph.D. Candidate, Biostatistician
Dept: Public Health Sciences
Div: Biostatistics
Collaboration and Consulting Core (BCCC)

It is common to meet high-dimensional data in modern data analysis. The higher dimensionality can dramatically degrade the performance of machine learning algorithm, which is known as the “curse of dimensionality”. Therefore, many dimensionality reduction techniques are developed to tackle the problem.

We divide these techniques into feature selection and feature extraction methods and introduce some examples in each class. Methods like LASSO, PCA, LDA, LLE and Autoencoder are mentioned. The presentation aims to give the audience a taste to those most commonly used dimensionality reduction methods.

RSVP Today!

After registering, you will receive a confirmation email containing information about joining the meeting. Any questions, please contact Maria Jimenez-Rodriguez.