DocumentCode
2872155
Title
Linear Regression Methods to Predict Interval-Valued Data
Author
Neto, Eufrasio A.Lima ; Carvalho, Francisco de A.T.de ; Bezerra, Lucas X T
Author_Institution
Centro de Informatica - CIn / UFPE, Cidade Universitaria, Brazil
fYear
2006
fDate
23-27 Oct. 2006
Firstpage
125
Lastpage
130
Abstract
This paper introduce a new criterion and two new linear regression methods to predict interval-valued data. The proposed approaches consist in a new point of view to study the relationship between the midpoints and the ranges of the interval-valued variables. The evaluation of the proposed prediction methods is based on the average behaviour of the root mean squared error and the square of the correlation coefficient in the framework of a Monte Carlo experiment in comparison with the method proposed by [3].
Keywords
Data analysis; Least squares approximation; Linear regression; Minimization methods; Monte Carlo methods; Neural networks; Prediction methods; Predictive models; Upper bound; Vectors;
fLanguage
English
Publisher
ieee
Conference_Titel
Neural Networks, 2006. SBRN '06. Ninth Brazilian Symposium on
Conference_Location
Ribeirao Preto, Brazil
Print_ISBN
0-7695-2680-2
Type
conf
DOI
10.1109/SBRN.2006.27
Filename
4026822
Link To Document