Dataset Information

A feature selection method based on multiple kernel learning with expression profiles of different types.

ABSTRACT: BACKGROUND:With the development of high-throughput technology, the researchers can acquire large number of expression data with different types from several public databases. Because most of these data have small number of samples and hundreds or thousands features, how to extract informative features from expression data effectively and robustly using feature selection technique is challenging and crucial. So far, a mass of many feature selection approaches have been proposed and applied to analyse expression data of different types. However, most of these methods only are limited to measure the performances on one single type of expression data by accuracy or error rate of classification. RESULTS:In this article, we propose a hybrid feature selection method based on Multiple Kernel Learning (MKL) and evaluate the performance on expression datasets of different types. Firstly, the relevance between features and classifying samples is measured by using the optimizing function of MKL. In this step, an iterative gradient descent process is used to perform the optimization both on the parameters of Support Vector Machine (SVM) and kernel confidence. Then, a set of relevant features is selected by sorting the optimizing function of each feature. Furthermore, we apply an embedded scheme of forward selection to detect the compact feature subsets from the relevant feature set. CONCLUSIONS:We not only compare the classification accuracy with other methods, but also compare the stability, similarity and consistency of different algorithms. The proposed method has a satisfactory capability of feature selection for analysing expression datasets of different types using different performance measurements.

SUBMITTER: Du W

PROVIDER: S-EPMC5288949 | biostudies-literature | 2017

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

A feature selection method based on multiple kernel learning with expression profiles of different types.

Du Wei W Cao Zhongbo Z Song Tianci T Li Ying Y Liang Yanchun Y

BioData mining 20170202

<h4>Background</h4>With the development of high-throughput technology, the researchers can acquire large number of expression data with different types from several public databases. Because most of these data have small number of samples and hundreds or thousands features, how to extract informative features from expression data effectively and robustly using feature selection technique is challenging and crucial. So far, a mass of many feature selection approaches have been proposed and applie ...[more]

PMID: 28184251

Similar Datasets

Project description:Introduction: DSM-5 presented a revised conceptualization of specific learning disorders (LD). Contrary to former versions, the various types of LD-i.e., mathematics disorder, reading disorder, and writing disorder-are not treated as distinct diagnostic entities but are integrated into one single LD category. In support of this new classification, it has been argued that the various types of LD overlap to a great extent in their cognitive functioning profiles and therefore do not exhibit a distinct set of cognitive causes. In contrast, ICD-11 still adheres to the idea of discrete categories and thus follows the specificity hypothesis of LD. Using latent profile analysis (LPA), we therefore tested the specificity of cognitive strengths and weaknesses in children with different types of LD. Secondly, we aimed at examining the extent to which observed LD characteristics (type and severity of LD as well as IQ-achievement discrepancy) were consistent with the membership of a given latent profile. Method: 302 German third-graders (134 girls; IQ ≥ 85; M age = 111.05 months; SD = 5.76) with single or comorbid types of LD in the domains of mathematics, reading, and spelling completed a wide range of domain-specific and domain-general cognitive functioning measures. Results: Five qualitative distinct profiles of cognitive strengths and weaknesses were identified. Profile 1 (23% of the sample) showed Comprehensive Cognitive Deficits, performing low in all measures except for naming speed, language, and inhibition. Profile 2 (21%) included children with a Double Deficit in Phonological Awareness and Phonological Short-term Memory. Profile 3 (20%) was characterized by a Double Deficit of Phonological Awareness and Naming Speed. Profile 4 (19%) included children with a Single Deficit in Attention, and profile 5 (17%) consisted of children without any cognitive deficits. Moreover, type and severity of LD as well as IQ-achievement discrepancy discriminated between the profiles, which is in line with the specificity hypothesis of LD. Discussion: Overall, the finding of specific associations between the LD types and the identified cognitive profiles supports the ICD-11 classification of LD. Yet, those inferences may not be valid for an individual child but need to be examined through comprehensive diagnostic.

Dataset Information

A feature selection method based on multiple kernel learning with expression profiles of different types.

Publications

A feature selection method based on multiple kernel learning with expression profiles of different types.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets