Erinija Pranckeviciene

Learn More
Many real-world classification problems are represented by very sparse and high-dimensional data. The recent successes of a linear programming support vector machine (LPSVM) for feature selection motivated a deeper analysis of the method when applied to sparse, multivariate data. Due to the sparseness, the selection of a classification model is greatly(More)
OBJECTIVE Demonstrate that incorporating domain knowledge into feature selection methods helps identify interpretable features with predictive capability comparable to a state-of-the-art classifier. METHODS Two feature selection methods, one using a genetic algorithm (GA) the other a L(1)-norm support vector machine (SVM), were investigated on three(More)
Every next generation sequencing (NGS) platform relies on proprietary and open source computational tools to analyze sequencing data. NGS tools for Illumina platforms are well documented which is not the case with AB SOLiD systems. We applied several computational and variant calling pipelines to analyse targeted exome sequencing data obtained using AB(More)
BACKGROUND Blepharophimosis is a fixed reduction in the vertical distance between the upper and lower eyelids with short palpebral fissures. It is a rare facial malformation and is considered an important diagnostic feature in dysmorphic analysis. It is likely that many patients with blepharophimosis-mental retardation syndrome have submicroscopic(More)
We investigated the geometrical complexity of several high-dimensional, small sample classification problems and its changes due to two popular feature selection procedures, forward feature selection (FFS) and Linear Programming Support Vector Machine (LPSVM). We found that both procedures are able to transform the problems to spaces of very low(More)
We investigate the relative efficacy of several classification models with and without feature selection. Simple classification rules are frequently preferable and superior to more complex models for microarray data that are typically undersampled. Improved classification accuracy is obtained with feature selection. We summarize some of the important(More)
ACE (I/D), ACTN3 (R/X), PPARGC1A (Gly482Ser) and PPARA (G/C) polymorphisms have been linked to the success in power-oriented sports through the intermediate phenotypes. The study involved 193 Lithuanian elite athletes and 250 controls. The measured phenotypic variables included short-term explosive muscle power (STEMP) and anaerobic alactic maximum power(More)
The identification of spectral signatures is crucial for the classification/profiling of biomedical spectra. Because only limited number of biomedical samples of high dimensionality is typically available, dimensionality reduction techniques (identification of discriminatory features) are essential for robust classifier development. We show, on three(More)