Lossless Pruning of AHP SNP Prioritizaton Tree Using Random Forest Variable Importances

Date

2017-06-30

Author

Yılmaz, Arif
Aydın Son, Yeşim

Metadata

Show full item record

Item Usage Stats

111
views

0
downloads

Subjectivity is an old yet unsolved problem in Multiple Criteria Decision Making including Analytic Hierarchy Processing (AHP). Here, we have proposed a machine learning based analytic hierarchy process (ML-AHP) method to address expert judgment uncertainty in decision making system design. It is accomplished by training a classifier algorithm according to Analytic Hierarchy Process input data and evaluating categories based on variable importance values. As a comparative case study Single Nucleotide Polymorphism(SNP) prioritization in the bioinformatics domain is presented. Variable Importance figures provided by the employed machine learner are used to evaluate the importanceof AHP categories. In this study we selected Random Forest. It was discovered that most of the expert defined weights of categories were zero and performance was identical after pruning. Analysis on the Prostate Cancer and the Type 2 Diabetes Mellitus disease data were performed to demonstrate the benefits of the proposed approach, where pairwise comparisons of categories required no expert evaluation. Hence, subjectivity, uncertainty and imprecision is avoided. Implementation of the proposed method can enhance evaluation of the category weights in AHP and also in other multiple criteria decision making methods

URI

http://hibit2017.ii.metu.edu.tr/wordpress/wp-content/uploads/HIBIT2017_Conference_Book.pdf
https://hdl.handle.net/11511/86284

Conference Name

10th International Symposium on Health Informatics and Bioinformatics

Collections

Graduate School of Informatics, Conference / Seminar

Suggestions

OpenMETU
Core

Fuzzy versus statistical linear regression Kim, KJ; Moskowitz, H; Köksalan, Mustafa Murat (1996-07-19) Statistical linear regression and fuzzy linear regression have been developed from different perspectives, and thus there exist several conceptual and methodological differences between the two approaches. The characteristics of both methods, in terms of basic assumptions, parameter estimation, and application are described and contrasted. Their descriptive and predictive capabilities are also compared via a simulation experiment to identify the conditions under which one outperforms the other. It turns out...
Stochastic discontinuous Galerkin methods for robust deterministic control of convection-diffusion equations with uncertain coefficients Çiloğlu, Pelin; Yücel, Hamdullah (2023-04-01) We investigate a numerical behavior of robust deterministic optimal control problem subject to a convection-diffusion equation containing uncertain inputs. Stochastic Galerkin approach, turning the original optimization problem containing uncertainties into a large system of deterministic problems, is applied to discretize the stochastic domain, while a discontinuous Galerkin method is preferred for the spatial discretization due to its better convergence behavior for optimization problems governed by conve...
A Favorable Weight-Based Evolutionary Algorithm for Multiple Criteria Problems SOYLU, Banu; Köksalan, Mustafa Murat (Institute of Electrical and Electronics Engineers (IEEE), 2010-04-01) In this paper, we present a favorable weight-based evolutionary algorithm for multiple criteria problems. The algorithm tries to both approximate the Pareto frontier and evenly distribute the solutions over the frontier. These two goals are common for many multiobjective evolutionary algorithms. To achieve these goals in our algorithm, each member selects its own weights for a weighted Tchebycheff distance function to define its fitness score. The fitness scores favor solutions that are closer to the Pareto...
Variable selection in robust heteroscedastic models with autoregressive covariance structures using EM-type algorithm Gökalp Yavuz, Fulya; GÜNEY, YEŞİM; ARSLAN, OLÇAY (2022-12-19) The joint modeling of location and scatter matrix with multivariate t-distribution provides a valuable extension to the classical approach with normal distribution when the data set under consideration involves outliers or heavy tail outcomes. Variable selection is essential in these models since the covariance model has three models built into it, and the number of unknown parameters grows quadratically with the size of the matrix. The first purpose is to obtain the maximum likelihood estimates of the para...
Bayesian learning under nonnormality Yılmaz, Yıldız Elif; Alpaslan, Ferda Nur; Department of Computer Engineering (2004) Naive Bayes classifier and maximum likelihood hypotheses in Bayesian learning are considered when the errors have non-normal distribution. For location and scale parameters, efficient and robust estimators that are obtained by using the modified maximum likelihood estimation (MML) technique are used. In naive Bayes classifier, the error distributions from class to class and from feature to feature are assumed to be non-identical and Generalized Secant Hyperbolic (GSH) and Generalized Logistic (GL) distribut...

Citation Formats

A. Yılmaz and Y. Aydın Son, “Lossless Pruning of AHP SNP Prioritizaton Tree Using Random Forest Variable Importances,” presented at the 10th International Symposium on Health Informatics and Bioinformatics, Güzelyurt, Kıbrıs, 2017, Accessed: 00, 2021. [Online]. Available: http://hibit2017.ii.metu.edu.tr/wordpress/wp-content/uploads/HIBIT2017_Conference_Book.pdf.