Positive and unlabeled learning involves positive examples and unlabeled data. The unlabeled data can contain both positive and negative examples. PU learning has gained prevalence recently due to its newfound application in social media and medicine. The current state of the art approaches to PU algorithms face a multitude of issues. Therefore the team implemented and conducted experiments on existing algorithms such as SAR-EM and NNPU. These algorithms were modified to create a novel PU algorithm.

• 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

• Gass, Bryan A.
• Abeyta, Jesse B.
• Cheng, Nicholas W.
• Nair, Vinay
• Kocienda, Calvin A.

Subject

• Computing
• Innovation

Publisher

• Worcester Polytechnic Institute

Identifier

• E-project-040621-212540
• 17241

Parola chiave

• Machine Learning
• Positive Unlabeled

Advisor

• Rundensteiner, Elke A.

Year

• 2021

Date created

• 2021-04-06

Resource type

• Major Qualifying Project

Major

• Computer Science

Rights statement

• In Copyright

Ultima modifica

• 2021-05-03