Informative Variables Selection for Multi-relational Supervised Learning

2011 ◽  
pp. 75-87 ◽  
Author(s):  
Dhafer Lahbib ◽  
Marc Boullé ◽  
Dominique Laurent
Keyword(s):  
Supervised Learning ◽  
Variables Selection ◽  
Selection For

scholarly journals MEASURING BANK EFFICIENCY: DEA APPLICATION

2014 ◽  
Vol 20 (4) ◽  
pp. 739-757 ◽  
Author(s):  
Jelena Titko ◽  
Jelena Stankevičienė ◽  
Nataļja Lāce
Keyword(s):  
Banking Sector ◽  
Linear Regression Analysis ◽  
Analytical Data ◽  
Returns To Scale ◽  
Mean Values ◽  
Dea Model ◽  
Two Samples ◽  
Variables Selection ◽  
Measuring Efficiency ◽  
Selection For

The paper aims to improve the methodology of measuring efficiency of Latvian banks. Efficiency scores were calculated with application of non-parametric frontier technique Data Envelopment Analysis (DEA). Input-oriented DEA model under Variable Returns to Scale (VRS) assumption was used. Potential model variables were selected based on the intermediation and profitability approach. Fourteen alternative models with different inputs-outputs combinations were developed for the research purposes. To substantiate the variables selection for DEA model the received data was processed, using such methods, as correlation analysis, linear regression analysis, analysis of mean values, and two-samples Kolmogorov-Smirnov test. The research results assisted the authors in providing general recommendations about the variables selection for DEA application in the Latvian banking sector. The present research contributes to the existing analytical data on bank performance in Latvia. The empirical findings provide a background for further studies, in particular, the efficiency of Latvian banks could be analysed in the extended time period.

scholarly journals Supervised learning-based tagSNP selection for genome-wide disease classifications

BMC Genomics ◽  
2008 ◽  
Vol 9 (Suppl 1) ◽  
pp. S6 ◽  
Author(s):  
Qingzhong Liu ◽  
Jack Yang ◽  
Zhongxue Chen ◽  
Mary Qu Yang ◽  
Andrew H Sung ◽  
...  
Keyword(s):  
Supervised Learning ◽  
Genome Wide ◽  
Selection For

Temperature Variables Selection for Thermal Error Modeling of Heavy-Duty CNC Milling-Boring Machine Tools

2011 ◽  
Vol 188 ◽  
pp. 171-174
Author(s):  
Gang Wei Cui ◽  
D. Gao ◽  
L. Wang ◽  
Y.X. Yao
Keyword(s):  
Machine Tools ◽  
Thermal Error ◽  
Error Modeling ◽  
Cnc Milling ◽  
Heavy Duty ◽  
Thermal Error Modeling ◽  
Selection Strategies ◽  
Variables Selection ◽  
Selection For ◽  
Feed Axis

One of the difficult issues in thermal error modeling is to select appropriate temperature variables. In this paper, two selection strategies are introduced to overcome this difficulty. After measuring the temperatures and thermal errors of a heavy-duty CNC milling-boring machine tool by a laser tracker, four temperature variables which are the foundation of thermal error modeling are selected for each feed axis from fifteen temperature variables according to major factor strategy and non-interrelated strategy.

On feature selection for supervised learning problems involving high-dimensional analytical information

RSC Advances ◽  
2016 ◽  
Vol 6 (86) ◽  
pp. 82801-82809 ◽  
Author(s):  
P. Žuvela ◽  
J. Jay Liu
Keyword(s):  
Feature Selection ◽  
Supervised Learning ◽  
Learning Problems ◽  
High Dimensional ◽  
Analytical Information ◽  
Selection For

Feature selection for supervised learning problems involving analytical information.

scholarly journals APPLICATION OF BAYESIAN FE MODEL SELECTION FOR SHM OF A MONUMENTAL MASONRY PALACE THROUGH SEMI-SUPERVISED LEARNING

2021 ◽  
Author(s):  
Laura Ierimonti ◽  
Nicola Cavalagli ◽  
Enrique García-Macías ◽  
Ilaria Venanzi ◽  
Filippo Ubertini
Keyword(s):  
Model Selection ◽  
Supervised Learning ◽  
Fe Model ◽  
Selection For
Sign in / Sign up

Export Citation Format

Share Document