A Survey of Evolutionary Algorithms for Decision-Tree Induction

Abstract

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.

Addressing the Class Imbalance Problem in Medical Datasets

“…K-Means [13] clustering is used to cluster the majority samples. For choosing the best subset, decision tree [11] and Fuzzy Unordered Rule Induction Algorithm [12]were used as classifiers. SM OTE is used as an oversampling technique.…”

“…All the subsets of the majority class are separately combined with the minority class samples to make K different training data sets (The K value is dependent on the data domain, in our implementation the K value was 3). All the combined datasets are classified using a decision tree [11] and a Fuzzy Unordered Rule Induction Algorithm [12]. The datasets with the highest accuracy over the majority of the classifiers were kept for further data mining processes.…”

A hyper-heuristic evolutionary algorithm for automatically designing decision-tree algorithms

“…Different strategies were employed for deriving accurate DTs, such as bottom-up induction [2], hybrid induction [22], evolutionary induction [1,3] and ensemble of trees [4], just to name a few. Nevertheless, no strategy has been more successful in generating accurate and comprehensible decision trees with low computational effort than the greedy top-down induction strategy.…”

From evolutionary computation to the evolution of things

“…For example, neuroevolutionary approaches use EAs to optimise the structure, the parameters, or both simultaneously, of artificial neural networks 33,34 . In other branches of machine learning, using EC to design algorithms has been shown to be very effective as an alternative to hand-crafting them, for instance, for inducing decision-trees 35 . Furthermore, EAs have been applied to prediction problems.…”