Feature selection and classification in noisy epistatic problems using a hybrid evolutionary approach

DeHaas, Drew; Craig, Jesse C.; Rickert, Colin; Eppstein, Margaret J.; Haake, P.; Stor, Kirsten

doi:10.1145/1276958.1277331

Cited by 3 publications

(3 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…All offspring derived from mutation are given the same age as the parent. DNFs selected for crossover undergo either set union or set intersection (Pru) [7]. The second parent is randomly selected, using tournament selection with replacement and a tournament size of 3.…”

Section: Evolving Dnf Rulesmentioning

confidence: 99%

Evolving Probabilistically Significant Epistatic Classification Rules for Heterogeneous Big Datasets

Hanley

Eppstein

Buzas

et al. 2016

Proceedings of the Genetic and Evolutionary Computation Conference 2016

Self Cite

View full text Add to dashboard Cite

We develop an algorithm to evolve sets of probabilistically significant multivariate feature interactions, with co-evolved feature ranges, for classification in large, complex datasets. The datasets may include nominal, ordinal, and/or continuous features, missing data, imbalanced classes, and other complexities. Our agelayered evolutionary algorithm generates conjunctive clauses to model multivariate interactions in datasets that are too large to be analyzed using traditional methods such as logistic regression. Using a novel hypergeometric probability mass function for fitness evaluation, the algorithm automatically archives conjunctive clauses that are probabilistically significant at a given threshold, thus identifying strong complex multivariate interactions. The method is validated on two synthetic epistatic datasets and applied to a complex real-world survey dataset aimed at determining the drivers of household infestation for an insect that transmits Chagas disease. We identify a set of 178,719 predictive feature interactions that are associated with household infestation, thus dramatically reducing the size of the search space for future analysis.

show abstract

Section: Evolving Dnf Rulesmentioning

confidence: 99%

Evolving Probabilistically Significant Epistatic Classification Rules for Heterogeneous Big Datasets

Hanley

Eppstein

Buzas

et al. 2016

Proceedings of the Genetic and Evolutionary Computation Conference 2016

Self Cite

View full text Add to dashboard Cite

show abstract

“…We implemented an intersection/union type of crossover, which previous work showed to be effective in feature selection applications [4]. Specifically, to cross two parents, we applied a bitwise AND operator 95% of the time (which returns the intersection of the feature sets of the two parents) and a bitwise OR operator 5% of the time (to help restore lost features) (see Figure 1).…”

Section: Feature Selection With a Gamentioning

confidence: 99%

“…Consequently, many researchers have successfully applied genetic algorithms (GAs) [9][20] [23][24] [26] [33] as feature selectors. In [4], a GA using a novel crossover method combining set intersection and set union was found to be very effective in identifying small non-linearly interacting sets of features in noisy data sets.…”

Section: Introductionmentioning

confidence: 99%

Evolutionary feature selection for classification

Krupa

Chatterjee

Eldridge

et al. 2012

Proceedings of the 14th Annual Conference on Genetic and Evolutionary Computation

Self Cite

View full text Add to dashboard Cite

We present a real-world application utilizing a Genetic Algorithm (GA) for exploratory multivariate association analysis of a large consumer survey designed to assess potential consumer adoption of Plug-in Hybrid Electric Vehicles (PHEVs). The GA utilizes an intersection/union crossover operator, in conjunction with high background mutation rates, to achieve rapid multivariate feature selection. We experimented with two alternative fitness measures based on classification results of a naïve Bayes quadratic discriminant analysis; one fitness function rewarded only for correct classifications, and the other penalized for the degree of misclassification using a quadratic penalty function. We achieved high classification accuracy for three different survey outcome questions (with 3-, 5-, and 7-outcome classes, respectively). The quadratic penalty function yielded better overall results, returning smaller feature sets and overall more accurate contingency tables of predicted classes. Our results help to identify what consumer attributes best predict their likelihood of purchasing a PHEV. These findings will be used to better inform an existing agent-based model of PHEV market penetration, with the ultimate aim of helping auto manufacturers and policy makers identify leverage points in the system that will encourage PHEV market adoption.

show abstract

Machine Learning in Mental Health

Thieme

Belgrave

Doherty

2020

ACM Trans. Comput.-Hum. Interact.

231

View full text Add to dashboard Cite

High prevalence of mental illness and the need for effective mental health care, combined with recent advances in AI, has led to an increase in explorations of how the field of machine learning (ML) can assist in the detection, diagnosis and treatment of mental health problems. ML techniques can potentially offer new routes for learning patterns of human behavior; identifying mental health symptoms and risk factors; developing predictions about disease progression; and personalizing and optimizing therapies. Despite the potential opportunities for using ML within mental health, this is an emerging research area, and the development of effective ML-enabled applications that are implementable in practice is bound up with an array of complex, interwoven challenges. Aiming to guide future research and identify new directions for advancing development in this important domain, this article presents an introduction to, and a systematic review of, current ML work regarding psycho-socially based mental health conditions from the computing and HCI literature. A quantitative synthesis and qualitative narrative review of 54 papers that were included in the analysis surfaced common trends, gaps, and challenges in this space. Discussing our findings, we (i) reflect on the current state-of-the-art of ML work for mental health, (ii) provide concrete suggestions for a stronger integration of human-centered and multidisciplinary approaches in research and development, and (iii) invite more consideration of the potentially far-reaching personal, social, and ethical implications that ML models and interventions can have, if they are to find widespread, successful adoption in real-world mental health contexts.

show abstract

Feature selection and classification in noisy epistatic problems using a hybrid evolutionary approach

Cited by 3 publications

References 20 publications

Evolving Probabilistically Significant Epistatic Classification Rules for Heterogeneous Big Datasets

Evolving Probabilistically Significant Epistatic Classification Rules for Heterogeneous Big Datasets

Evolutionary feature selection for classification

Machine Learning in Mental Health

Contact Info

Product

Resources

About