TPOT: A Tree-based Pipeline Optimization Tool for Automating Machine Learning

Randal S. Olson, Jason H. Moore
Proceedings of the Workshop on Automatic Machine Learning, PMLR 64:66-74, 2016.

Abstract

As data science becomes more mainstream, there will be an ever-growing demand for data science tools that are more accessible, flexible, and scalable. In response to this demand, automated machine learning (autoML) researchers have begun building systems that automate the process of designing and optimizing machine learning pipelines. In this paper we present TPOT, an open source genetic programming-based autoML system that optimizes a series of feature preprocessors and machine learning models with the goal of maximizing classification accuracy on a supervised classification task. We benchmark TPOT on a series of 150 supervised classification tasks and find that it significantly outperforms a basic machine learning analysis in 22 of them, while experiencing minimal degradation in accuracy on 5 of the benchmarks—all without any domain knowledge nor human input. As such, GP-based autoML systems show considerable promise in the autoML domain.

Cite this Paper


BibTeX
@InProceedings{pmlr-v64-olson_tpot_2016, title = {TPOT: A Tree-based Pipeline Optimization Tool for Automating Machine Learning}, author = {Olson, Randal S. and Moore, Jason H.}, booktitle = {Proceedings of the Workshop on Automatic Machine Learning}, pages = {66--74}, year = {2016}, editor = {Hutter, Frank and Kotthoff, Lars and Vanschoren, Joaquin}, volume = {64}, series = {Proceedings of Machine Learning Research}, address = {New York, New York, USA}, month = {24 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v64/olson_tpot_2016.pdf}, url = {https://proceedings.mlr.press/v64/olson_tpot_2016.html}, abstract = {As data science becomes more mainstream, there will be an ever-growing demand for data science tools that are more accessible, flexible, and scalable. In response to this demand, automated machine learning (autoML) researchers have begun building systems that automate the process of designing and optimizing machine learning pipelines. In this paper we present TPOT, an open source genetic programming-based autoML system that optimizes a series of feature preprocessors and machine learning models with the goal of maximizing classification accuracy on a supervised classification task. We benchmark TPOT on a series of 150 supervised classification tasks and find that it significantly outperforms a basic machine learning analysis in 22 of them, while experiencing minimal degradation in accuracy on 5 of the benchmarks—all without any domain knowledge nor human input. As such, GP-based autoML systems show considerable promise in the autoML domain.} }
Endnote
%0 Conference Paper %T TPOT: A Tree-based Pipeline Optimization Tool for Automating Machine Learning %A Randal S. Olson %A Jason H. Moore %B Proceedings of the Workshop on Automatic Machine Learning %C Proceedings of Machine Learning Research %D 2016 %E Frank Hutter %E Lars Kotthoff %E Joaquin Vanschoren %F pmlr-v64-olson_tpot_2016 %I PMLR %P 66--74 %U https://proceedings.mlr.press/v64/olson_tpot_2016.html %V 64 %X As data science becomes more mainstream, there will be an ever-growing demand for data science tools that are more accessible, flexible, and scalable. In response to this demand, automated machine learning (autoML) researchers have begun building systems that automate the process of designing and optimizing machine learning pipelines. In this paper we present TPOT, an open source genetic programming-based autoML system that optimizes a series of feature preprocessors and machine learning models with the goal of maximizing classification accuracy on a supervised classification task. We benchmark TPOT on a series of 150 supervised classification tasks and find that it significantly outperforms a basic machine learning analysis in 22 of them, while experiencing minimal degradation in accuracy on 5 of the benchmarks—all without any domain knowledge nor human input. As such, GP-based autoML systems show considerable promise in the autoML domain.
RIS
TY - CPAPER TI - TPOT: A Tree-based Pipeline Optimization Tool for Automating Machine Learning AU - Randal S. Olson AU - Jason H. Moore BT - Proceedings of the Workshop on Automatic Machine Learning DA - 2016/12/04 ED - Frank Hutter ED - Lars Kotthoff ED - Joaquin Vanschoren ID - pmlr-v64-olson_tpot_2016 PB - PMLR DP - Proceedings of Machine Learning Research VL - 64 SP - 66 EP - 74 L1 - http://proceedings.mlr.press/v64/olson_tpot_2016.pdf UR - https://proceedings.mlr.press/v64/olson_tpot_2016.html AB - As data science becomes more mainstream, there will be an ever-growing demand for data science tools that are more accessible, flexible, and scalable. In response to this demand, automated machine learning (autoML) researchers have begun building systems that automate the process of designing and optimizing machine learning pipelines. In this paper we present TPOT, an open source genetic programming-based autoML system that optimizes a series of feature preprocessors and machine learning models with the goal of maximizing classification accuracy on a supervised classification task. We benchmark TPOT on a series of 150 supervised classification tasks and find that it significantly outperforms a basic machine learning analysis in 22 of them, while experiencing minimal degradation in accuracy on 5 of the benchmarks—all without any domain knowledge nor human input. As such, GP-based autoML systems show considerable promise in the autoML domain. ER -
APA
Olson, R.S. & Moore, J.H.. (2016). TPOT: A Tree-based Pipeline Optimization Tool for Automating Machine Learning. Proceedings of the Workshop on Automatic Machine Learning, in Proceedings of Machine Learning Research 64:66-74 Available from https://proceedings.mlr.press/v64/olson_tpot_2016.html.

Related Material