Discovering and understanding change using multivariate trees: the RECPAM approach

Authors:
Antonio Ciampi;Alina Dyachenko
Affiliations:
Department of Epidemiology and Biostatistics, McGill University, Montreal, QC, Canada;Department of Epidemiology and Biostatistics, McGill University, Montreal, QC, Canada
Venue:
MATH'05 Proceedings of the 8th WSEAS International Conference on Applied Mathematics
Year:
2005

Citing 5
Cited 0

Generalized regression trees

Computational Statistics & Data Analysis - Data analysis and inference in nonstandard settings
Principles of data mining

Principles of data mining
Bayesian Treed Models

Machine Learning
Induction of Decision Trees

Machine Learning
Tree-structured subgroup analysis for censored survival data: Validation of computationally inexpensive model selection criteria

Statistics and Computing

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper was developed to understand a particular problem through data mining. A web-based questionnaire is used to allow visitors to evaluate a particular website and provide feedback to the owner. The questionnaire is proposed to the user over two time windows. We are interested in the following questions. Does the average evaluation change over time? Is this change uniform or does it vary across special visitor subgroups? If there are special subgroups, can we describe them? An answer to these questions is offered by tree-structured data mining, having as target the coefficient of the time variable of a multivariate predictor. The proposed approach is applicable to the general problem of detecting and understanding change when a flow of data is observed through several time windows.