In this project, we applied the deep learning methods of autoencoder-Kalman filtering as well as the autoencoder preprocessing from Ciecierski [3] to improve clustering on action potentials by filtering noise. We used multiple types of clustering, including k-means clustering, mean-shift clustering, and agglomerative hierarchical clustering. We evaluated the performance of each clustering algorithm after using each filtering method, as well as no filtering, to analyze which of the filtering methods has the best effect on clustering action potentials.

• This report represents the work of one or more WPI undergraduate students submitted to the faculty as evidence of completion of a degree requirement. WPI routinely publishes these reports on its website without editorial or peer review.

Creator

• Reimert Burro, Erik
• Cleaver-Stigum, Imogen

Publisher

• Worcester Polytechnic Institute

Identifier

• 17236
• E-project-040621-180932

Keyword

• deep learning
• artificial intelligence
• machine learning
• clustering
• autoencoder
• signal processing

Advisor

• Smith, Therese Mary

Year

• 2021

Date created

• 2021-04-06

Resource type

• Major Qualifying Project

Major

• Computer Science

Rights statement

• In Copyright

License

• All rights reserved

Last modified

• 2021-05-03