University of Illinois Urbana-Champaign

Structured machine learning: A language-agnostic framework for streamlining the development of machine learning pipelines

Hao, Michael

This item is only available for download by members of the University of Illinois community. Students, faculty, and staff at the U of I may log in with your NetID and password to view the item. If you are trying to access an Illinois-restricted dissertation or thesis, you can request a copy through your library's Inter-Library Loan office or purchase a copy directly from ProQuest.

Permalink

https://hdl.handle.net/2142/97858

Description

Title
Structured machine learning: A language-agnostic framework for streamlining the development of machine learning pipelines
Author(s)
Hao, Michael
Contributor(s)
Brunner, Robert
Issue Date
2017-05
Keyword(s)
  • Machine learning
  • Python
  • Scikit-learn
  • Pandas
  • BNF
Abstract
Standard machine learning (SML) is a proof of concept meta-language to explore the idea of abstracting the implementation details of machine learning frameworks into a simple, concise language. Having such a language presents various benefits such as the standardization of the machine learning software development “stack” to reproducible research in the machine learning academic community. The SML framework is developed in Python, which consists of an object-oriented BNF parser, a Python scikit-learn/pandas driver, and a connector that puts the parser in contact with the driver. Drivers were implemented for building regression, clustering, and classification models, as well as preprocessing actions such as simple data imputation. Ten popular machine learning problems were solved via SML to test the effectiveness of the prototype implementation. Solutions to difficult problems such as hyperparameter selection via the language as well as model persistence are still being explored.
Type of Resource
text
Language
en
Permalink
http://hdl.handle.net/2142/97858

Owning Collections

Senior Theses - Electrical and Computer Engineering PRIMARY
The best of ECE undergraduate research