Classification of genomic data: Some aspects of feature selection

Czekaj, Tomasz; Wu, Wen; Walczak, Beata

doi:10.1016/j.talanta.2008.03.045

Cited by 13 publications

(4 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Besides descriptor importances from decision tree, feature selection was conducted by selecting the best descriptors based on the UV statistical tests for dimensionality reduction purpose. Here, F-test was adopted to estimate the degree of linear dependency between descriptor and target 57,58 . Briefly, F-test of equality of variances is a test for the null hypothesis that two normal populations have the same variance 59 .…”

Section: Dataset Collectionmentioning

confidence: 99%

Descriptor selection for predicting interfacial thermal resistance by machine learning methods

Tian

Chen

2021

Sci Rep

View full text Add to dashboard Cite

Interfacial thermal resistance (ITR) is a critical property for the performance of nanostructured devices where phonon mean free paths are larger than the characteristic length scales. The affordable, accurate and reliable prediction of ITR is essential for material selection in thermal management. In this work, the state-of-the-art machine learning methods were employed to realize this. Descriptor selection was conducted to build robust models and provide guidelines on determining the most important characteristics for targets. Firstly, decision tree (DT) was adopted to calculate the descriptor importances. And descriptor subsets with topX highest importances were chosen (topX-DT, X = 20, 15, 10, 5) to build models. To verify the transferability of the descriptors picked by decision tree, models based on kernel ridge regression, Gaussian process regression and K-nearest neighbors were also evaluated. Afterwards, univariate selection (UV) was utilized to sort descriptors. Finally, the top5 common descriptors selected by DT and UV were used to build concise models. The performance of these refined models is comparable to models using all descriptors, which indicates the high accuracy and reliability of these selection methods. Our strategy results in concise machine learning models for a fast prediction of ITR for thermal management applications.

show abstract

Section: Dataset Collectionmentioning

confidence: 99%

Descriptor selection for predicting interfacial thermal resistance by machine learning methods

Tian

Chen

2021

Sci Rep

View full text Add to dashboard Cite

show abstract

“…The main aim of genomics and proteomics is to characterize and to dig out all the pharmacology related information using computational methods (21,22). However, as majority of proteins have no known function, therefore, this goal cannot be reached for all the proteins.…”

Section: Bioinformaticsmentioning

confidence: 99%

Pragmatic Approach To Drug Discovery Through Biophysical Perspective

Alok

Kumar²,

Kumar³

et al. 2016

IJMSCI

View full text Add to dashboard Cite

Drug discovery process is a long drawn time and resource consuming process. The conventional drug discovery process involves potential lead from either traditional system of medicine or from the in vitro experiments leading to a lot of animal experimentations wahich culminates in human trials. These whole processes usually take years of time and resources. Many a times, when a drug accreditation is accorded, the pharmacoeconomic values outnumbers the drug addition value to the drug library. With the advent of the usage of computers in biology, the discovery process shortened with the help of uniform and defined in silico parameters, thus providing an alternative to the laborious and variable in vitro experiments. While, this was a welcome move in the field of drug discovery but simultaneously it also opened the flood of potential molecules which often matched the defined in silico parameters but provided variable results in in vivo experiments. This article provides an overview of the drug discovery processes and its efficacy with a check on pharmacoeconomic values. The system of selection of potential lead molecule is based on several parameters out of which the most important aspect is that it should have no toxicity in vivo.

show abstract

“…Although the embedded variable selection process is well-known ( [8,15,20]), it may appear at first glance that it is not correctly handled in the optimization program…”

Section: Classification Trees and Variable Selectionmentioning

confidence: 99%

Risk Bounds for Embedded Variable Selection in Classification Trees

Mary‐Huard¹

2014

IEEE Trans. Inform. Theory

View full text Add to dashboard Cite

The problems of model and variable selections for classification trees are jointly considered. A penalized criterion is proposed which explicitly takes into account the number of variables, and a risk bound inequality is provided for the tree classifier minimizing this criterion. This penalized criterion is compared to the one used during the pruning step of the CART algorithm. It is shown that the two criteria are similar under some specific margin assumptions. In practice, the tuning parameter of the CART penalty has to be calibrated by hold-out. Simulation studies are performed which confirm that the hold-out procedure mimics the form of the proposed penalized criterion.

show abstract

Classification of genomic data: Some aspects of feature selection

Cited by 13 publications

References 19 publications

Descriptor selection for predicting interfacial thermal resistance by machine learning methods

Descriptor selection for predicting interfacial thermal resistance by machine learning methods

Pragmatic Approach To Drug Discovery Through Biophysical Perspective

Risk Bounds for Embedded Variable Selection in Classification Trees

Contact Info

Product

Resources

About