MissForest—non-parametric missing value imputation for mixed-type data

Bioinformatics (Oxford, England) - Tập 28 Số 1 - Trang 112-118 - 2012
Daniel J. Stekhoven1, Peter Bühlmann1
11 Seminar for Statistics, Department of Mathematics, ETH Zurich, 2Life Science Zurich PhD Program on Systems Biology of Complex Diseases and 3Competence Center for Systems Physiology and Metabolic Diseases, Zurich, Switzerland

Tóm tắt

Abstract Motivation: Modern data acquisition based on high-throughput technology is often facing the problem of missing data. Algorithms commonly used in the analysis of such large-scale data often depend on a complete set. Missing value imputation offers a solution to this problem. However, the majority of available imputation methods are restricted to one type of variable only: continuous or categorical. For mixed-type data, the different types are usually handled separately. Therefore, these methods ignore possible relations between variable types. We propose a non-parametric method which can cope with different types of variables simultaneously. Results: We compare several state of the art methods for the imputation of missing values. We propose and evaluate an iterative imputation method (missForest) based on a random forest. By averaging over many unpruned classification or regression trees, random forest intrinsically constitutes a multiple imputation scheme. Using the built-in out-of-bag error estimates of random forest, we are able to estimate the imputation error without the need of a test set. Evaluation is performed on multiple datasets coming from a diverse selection of biological fields with artificially introduced missing values ranging from 10% to 30%. We show that missForest can successfully handle missing values, particularly in datasets including different types of variables. In our comparative study, missForest outperforms other methods of imputation especially in data settings where complex interactions and non-linear relations are suspected. The out-of-bag imputation error estimates of missForest prove to be adequate in all settings. Additionally, missForest exhibits attractive computational efficiency and can cope with high-dimensional data. Availability: The ℝ package missForest is freely available from http://stat.ethz.ch/CRAN/. Contact:  [email protected]; [email protected]

Từ khóa


Tài liệu tham khảo

Breiman, 2001, Random forests, Mach. Learn., 45, 5, 10.1023/A:1010933404324

Frank, 2010, UCI machine learning repository.

Harley, 1987, Analysis of e. coli promoter sequences, Nucleic Acids Res., 15, 2343, 10.1093/nar/15.5.2343

Kurgan, 2001, Knowledge discovery approach to automated cardiac SPECT diagnosis, Artif. Intell. Med., 23, 149, 10.1016/S0933-3657(01)00082-3

Latal, 2009, Psychological adjustment and quality of life in children and adolescents following open-heart surgery for congenital heart disease: a systematic review, BMC Pediatr., 9, 6, 10.1186/1471-2431-9-6

Li, 1988, Imputation using Markov chains, J. Stat. Comput. Simul., 30, 57, 10.1080/00949658808811085

Little, 2008, Suitability of dysphonia measurements for telemonitoring of Parkinson's disease, Nature Precedings., 10.1038/npre.2008.2298.1

Little, 1987, Statistical Analysis with Missing Data.

Little, 1985, Maximum likelihood estimation for mixed continuous and categorical data with missing values, Biometrika, 72, 497, 10.1093/biomet/72.3.497

Mazumder, 2010, Spectral regularization algorithms for learning large incomplete matrices, J. Mach. Learn. Res., 11, 2287

Oba, 2003, A Bayesian missing value estimation method for gene expression profile data, Bioinformatics, 19, 2088, 10.1093/bioinformatics/btg287

Raghunathan, 2001, A multivariate technique for multiply imputing missing values using a sequence of regression models, Surv. Methodol., 27, 85

Rubin, 1978, Multiple imputations in sample surveys-a phenomenological Bayesian approach to nonresponse, Proceedings of the Survey Research Methods Section, American Statistical Association, 20

Rubin, 1990, Efficiently creating multiple imputations for incomplete multivariate normal data, Proceedings of the Statistical Computing Section of the American Statistical Association, 83

Schafer, 1997, Analysis of Incomplete Multivariate Data., 10.1201/9781439821862

Smit, 2007, Assessing the statistical validity of proteomics based biomarkers, Anal. Chim. Acta, 592, 210, 10.1016/j.aca.2007.04.043

Städler, 2010, Pattern alternating maximization algorithm for high-dimensional missing data.

Tibshirani, 1996, Regression shrinkage and selection via the lasso, J. R. Stat. Soc. Ser. B, 58, 267

Towell, 1990, Refinement of approximate domain theories by knowledge-based neural networks, Proceedings of the Eighth National Conference on Artificial Intelligence, 861

Troyanskaya, 2001, Missing value estimation methods for DNA microarrays, Bioinformatics, 17, 520, 10.1093/bioinformatics/17.6.520

Van Buuren, 2007, Multiple imputation of discrete and continuous data by fully conditional specification, Stat. Methods Med. Res., 16, 219, 10.1177/0962280206074463

Van Buuren, 1999, Flexible Multivariate Imputation by MICE.

Wille, 2004, Sparse graphical Gaussian modeling of the isoprenoid gene network in Arabidopsis thaliana, Genome Biol., 5, R92, 10.1186/gb-2004-5-11-r92

Wu, 2007, The effect of insulin on expression of genes and biochemical pathways in human skeletal muscle, Endocrine, 31, 5, 10.1007/s12020-007-0007-x

Thông tin
Thông tin xuất bản

Bioinformatics (Oxford, England)

Tập 28 Số 1

112-118

Thông tin tác giả