Sparse Omics-network Regularization to Increase Interpretability and Performance of SVM-based Predictive Models
The result's identifiers
Result code in IS VaVaI
<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F15%3A00234069" target="_blank" >RIV/68407700:21230/15:00234069 - isvavai.cz</a>
Result on the web
—
DOI - Digital Object Identifier
—
Alternative languages
Result language
angličtina
Original language name
Sparse Omics-network Regularization to Increase Interpretability and Performance of SVM-based Predictive Models
Original language description
To fully profit from development of high-throughput omics technologies, there is a strict need for accurate, stable and comprehensible biomarkers. The biomarkers are features of mostly molecular character, which enable to predict end interpret the individual?s state. However, the task of highthroughput data analysis is still challenging. Small sample size together large feature space often causes overfitting. Next, resulting model are difficult to interpret due to complex nature of omics processes. In this paper we propose a framework for effective implementation of large scale optimization problem within machine learning complex. The core algorithm is intended to improve SVM based linear models of gene expression as to the accuracy and especially thecomprehensibility. The algorithm, called SNSVM, uses regularization to achieve these objectives. The regularization is implemented through prior known feature interactions and additional sparsity term. The results suggest that prior knowl
Czech name
—
Czech description
—
Classification
Type
D - Article in proceedings
CEP classification
IN - Informatics
OECD FORD branch
—
Result continuities
Project
—
Continuities
S - Specificky vyzkum na vysokych skolach
Others
Publication year
2015
Confidentiality
S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů
Data specific for result type
Article name in the collection
Proceedings of the 19th International Scientific Student Conferenece POSTER 2015
ISBN
978-80-01-05499-4
ISSN
—
e-ISSN
—
Number of pages
1
Pages from-to
—
Publisher name
Czech Technical University in Prague
Place of publication
Praha
Event location
Praha
Event date
May 14, 2015
Type of event by nationality
EUR - Evropská akce
UT code for WoS article
—