Evolutionary model tree induction

Barros, Rodrigo C.; Basgalupp, Márcio P.; Ruiz, Duncan D.; Carvalho, André C. P. L. F. de; Freitas, Alex A.

doi:10.1145/1774088.1774327

Cited by 13 publications

(16 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…is the standard deviation, D is the portion of the instances that reaches the node being tested and Di is the portion of the instances that results from partitioning the node. This approach has two main advantages [4,6]: i) the threshold values are defined in a data-driven manner instead of selecting random values, which is the case of most evolutionary approaches; and ii) we can achieve a certain degree of heterogeneity by partitioning the training set (in this case, the imputed sub-training set) into different pieces, increasing the chances of selecting a good threshold value.…”

Section: Generating the Initial Populationmentioning

confidence: 99%

Evolving regression trees robust to missing data

Blomberg

Barros

Ruiz

2015

Proceedings of the 30th Annual ACM Symposium on Applied Computing

Self Cite

View full text Add to dashboard Cite

Data quality is a major concern in several fields of knowledge that rely on data analysis. Missing data, in particular, have a strong negative impact in machine learning, potentially harming the knowledge extraction process by skewing results and affecting the predictive performance of the induced models. For dealing with the problem of missing data, the literature in machine learning offers a variety of strategies which can be either in the form of a preprocessing step or of an embedded solution within a predictive method. In this paper, we propose a novel evolutionary algorithm for regression tree induction, which has embedded in its evolutionary cycle a robust framework for dealing with missing data. For comparison purposes, we evaluate six traditional regression algorithms over 10 public regression datasets that were artificially modified to present different levels of missing data. Results from the experimental analysis show that the proposed approach is the one that is less affected by the increasing levels of missing data, presenting an interesting trade-off between model interpretability and predictive performance especially for datasets with more than 40% of missing data.

show abstract

Section: Generating the Initial Populationmentioning

confidence: 99%

Evolving regression trees robust to missing data

Blomberg

Barros

Ruiz

2015

Proceedings of the 30th Annual ACM Symposium on Applied Computing

Self Cite

View full text Add to dashboard Cite

show abstract

“…Barros et al [46], [47] propose an EA called E-Motion (Evolutionary MOdel Tree InductiON) for axis-parallel model tree induction, where each individual is represented as a tree of variable shape and size.The initialization of individuals is domain knowledge-based, as it combines single nodes whose attribute tests are dictated by the expectation of standard deviation reduction (SDR), given by…”

Section: B Model Treesmentioning

confidence: 99%

“…Nevertheless, many works affirm to present similar predictive performance to baseline algorithms while providing smaller trees (e.g., [17], [46], [47], [100], [102], [103], [105], [111]). Only one work [120] states that the proposed EA generates larger trees than its baseline algorithm (C4.5).…”

Section: Performance Analysesmentioning

confidence: 99%

A Survey of Evolutionary Algorithms for Decision-Tree Induction

Barros

Basgalupp

Carvalho

et al. 2012

IEEE Trans. Syst., Man, Cybern. C

Self Cite

296

139

View full text Add to dashboard Cite

Abstract-This paper presents a survey of evolutionary algorithms designed for decision tree induction. In this context, most of the paper focuses on approaches that evolve decision trees as an alternate heuristics to the traditional top-down divideand-conquer approach. Additionally, we present some alternative methods that make use of evolutionary algorithms to improve particular components of decision tree classifiers. The paper original contributions are the following. First, it provides an upto-date overview that is fully focused on evolutionary algorithms and decision trees and does not concentrate on any specific evolutionary approach. Second, it provides a taxonomy which addresses works that evolve decision trees and works that design decision tree components using evolutionary algorithms. Finally, a number of references is provided that describe applications of evolutionary algorithms for decision tree induction in different domains. The paper ends by addressing some important issues and open questions that can be subject of future research.

show abstract

“…The decision trees produced by this strategy are usually fully balanced, where the distance from the root to any leaf node is the same. This generation procedure is called the full method (see, for instance, [71,66,72,68,44,6,7]). Trees can also be generated with varying distances from the root to the leaves, named the grow method (e.g., [48]).…”

Section: Related Workmentioning

confidence: 99%