Software | Machine Learning Modeling | Technology: Tools, Hardware, and Software | All Resources

Extended Supervised Tracking and Classification System (scikit-ExSTraCS)

by

The scikit-ExSTraCS package includes a sklearn-compatible Python implementation of ExSTraCS 2.0. ExSTraCS 2.0, or Extended Supervised Tracking and Classifying System, implements the core components of a Michigan-Style Learning Classifier System (where the system’s genetic algorithm operates on a rule level, evolving a population of rules with each their own parameters) in an easy to understand way, while still being highly functional in solving ML problems. It allows the incorporation of expert knowledge in the form of attribute weights, attribute tracking, rule compaction, and a rule specificity limit, that makes it particularly adept at solving highly complex problems. In general, Learning Classifier Systems (LCSs) are a classification of Rule Based Machine Learning Algorithms that have been shown to perform well on problems involving high amounts of heterogeneity and epistasis. Well designed LCSs are also highly human interpretable. LCS variants have been shown to adeptly handle supervised and reinforced, classification and regression, online and offline learning problems, as well as missing or unbalanced data. These characteristics of versatility and interpretability give LCSs a wide range of potential applications, notably those in biomedicine.

 

View Resource
Software | Feature learning | Technology: Tools, Hardware, and Software | All Resources

Multifactor Dimensionality Reduction (scikit-MDR)

by

A scikit-learn-compatible Python implementation of Multifactor Dimensionality Reduction (MDR) for feature construction. This project is still under active development and we encourage you to check back on this repository regularly for updates. MDR is an effective feature construction algorithm that is capable of modeling higher-order interactions and capturing complex patterns in data sets. MDR currently only works with categorical features and supports both binary classification and regression problems. We are working on expanding the algorithm to cover more problem types and provide more convenience features.

 

View Resource
Software | Technology: Tools, Hardware, and Software | Feature selection | All Resources

Relief-based Algorithm Training Environment (REBATE)

by

This package includes a scikit-learn-compatible Python implementation of ReBATE, a suite of Relief-based feature selection algorithms for Machine Learning. These Relief-Based algorithms (RBAs) are designed for feature weighting/selection as part of a machine learning pipeline (supervised learning). Presently this includes the following core RBAs: ReliefF, SURF, SURF*, MultiSURF*, and MultiSURF. Additionally, an implementation of the iterative TuRF mechanism and VLSRelief is included. It is still under active development and we encourage you to check back on this repository regularly for updates. These algorithms offer a computationally efficient way to perform feature selection that is sensitive to feature interactions as well as simple univariate associations, unlike most currently available filter-based feature selection methods. The main benefit of Relief algorithms is that they identify feature interactions without having to exhaustively check every pairwise interaction, thus taking significantly less time than exhaustive pairwise search.

 

View Resource
Software | AutoML | Technology: Tools, Hardware, and Software | All Resources

Simple, Transparent, End-to-end Automated Machine Learning Pipeline (STREAMLINE)

by

STREAMLINE is an end-to-end automated machine learning (AutoML) pipeline that empowers anyone to easily run, interpret, and apply a rigorous and customizable analysis for data mining or predictive modeling. Notably, this tool is currently limited to supervised learning on tabular, binary classification data but will be expanded as our development continues. The development of this pipeline focused on (1) overall automation, (2) avoiding and detecting sources of bias, (3) optimizing modeling performance, (4) ensuring complete reproducibility (under certain STREAMLINE parameter settings), (5) capturing complex associations in data (e.g. feature interactions), and (6) enhancing interpretability of output. Overall, the goal of this pipeline is to provide a transparent framework to learn from data as well as identify the strengths and weaknesses of ML modeling algorithms or other AutoML algorithms.

 

View Resource