Learning Decision Trees for Unbalanced Data

By David A. Cieslak, Nitesh V. Chawla. 2008

301 citations recorded by the source corpus.

Subfields: Accounting, Artificial Intelligence, Information Systems

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Accounting

In Artificial Intelligence

In Information Systems

Citations in the corpus

A Review on Ensembles for the Class Imbalance Problem: Bagging-, Boosting-, and Hybrid-Based Approaches (2012, methodology): “However, in [106], the authors show that it often experiences a reduction in performance when sampling techniques are applied, which is the base of the majority of the studied techniques; moreover, being more robust (less weak) than C4.”
FSVM-CIL: Fuzzy Support Vector Machines for Class Imbalance Learning (2010, methodology): “Algo-rithmic techniques have been developed for the different clas-siﬁcation algorithms, such as neural networks [26], decision trees [27], fuzzy systems [28], [29] etc., for imbalanced dataset learning.”
Learned lessons in credit card fraud detection from a practitioner perspective (2014, methodology): “HDDT [50] is a decision tree that uses Hellinger distance [51] as splitting criteria that is able to deal with skewed distribution.”
Class Confidence Weighted kNN Algorithms for Imbalanced Data Sets (2011, methodology): “For example, modiﬁcations of decision tree algorithms have been proposed to improve the standard C4.5, such as HDDT [4] and CCPDT [5].”
Hellinger distance decision trees are robust and skew-insensitive (2011, methodology): “In light of the observations within this report, we claim that HDDT are not only skew-insensitive as suggested in Cieslak and Chawla (2008a) , but also robust in their applicability to wide variety of datasets.”
Training and assessing classification rules with imbalanced data (2012, methodology): “This is the case, e.g., for classiﬁcation trees, discussed in the context of imbalance learning by Chawla (2003) and Cieslak and Chawla (2008) or for support vector machines (Akbani et al. 1999).”
Hybrid support vector machine rule extraction method for discovering the preferences of stock market investors: Evidence from Montenegro (2015, background): “…markets, where the number of annual transactions is too small and where the data are unbalanced, the performance of DT models is not satisfactory (learning from unbalanced datasets presents a convoluted problem in which traditional learning algorithms may perform poorly, Cieslak & Chawla, 2008).”
Adaptive Machine Learning for Credit Card Fraud Detection (2015, methodology): “1 Hellinger Distance Decision Trees Originally introduced to quantify the similarity between two probability distributions [233], the Hellinger Distance (HD) has been recently proposed as a splitting criterion in decision trees to improve the accuracy in unbalanced problems [103, 230].”
A Survey of Predictive Modeling on Imbalanced Domains (2016, methodology): “Hellinger distance was introduced as a decision tree splitting criterion to build Hellinger Distance Decision Trees (HDDT) [Cieslak and Chawla 2008].”
New perspectives and methods in link prediction (2010, methodology): “training skew-insensitive decision trees based on Hellinger distance [7].”