A class skew-insensitive ACO-based decision tree algorithm for imbalanced data sets

<span>Ant-tree-miner (ATM) has an advantage over the conventional decision tree algorithm in terms of feature selection. However, real world applications commonly involved imbalanced class problem where the classes have different importance. This condition impeded the entropy-based heuristic of existing ATM algorithm to develop effective decision boundaries due to its biasness towards the dominant class. Consequently, the induced decision trees are dominated by the majority class which lack in predictive ability on the rare class. This study proposed an enhanced algorithm called hellinger-ant-tree-miner (HATM) which is inspired by ant colony optimization (ACO) metaheuristic for imbalanced learning using decision tree classification algorithm. The proposed algorithm was compared to the existing algorithm, ATM in nine (9) publicly available imbalanced data sets. Simulation study reveals the superiority of HATM when the sample size increases with skewed class (Imbalanced Ratio &lt; 50%). Experimental results demonstrate the performance of the existing algorithm measured by BACC has been improved due to the class skew-insensitiveness of hellinger distance. The statistical significance test shows that HATM has higher mean BACC score than ATM.</span>

Subpopulation Discovery in Epidemiological Data with Subspace Clustering

A prerequisite of personalized medicine is the identification of groups of people who share specific risk factors towards an outcome. We investigate the potential of subspace clustering for finding such groups in epidemiological data. We propose a workflow that encompasses clusterability assessment before cluster discovery and quality assessment after learning the clusters. Epidemiological usually do not have a ground truth for the verification of clusters found in subspaces. Hence, we introduce quality assessment through juxtaposition of the learned models to “models-of-randomness”, i.e. models that do not reflect a true cluster structure. On the basis of this workflow, we select subspace clustering methods, compare and discuss their performance. We use a dataset with hepatic steatosis as outcome, but our findings apply on arbitrary epidemiological cohort data that have tenths of variables and exhibit class skew.

Model Assessment with ROC Curves

Classification models and in particular binary classification models are ubiquitous in many branches of science and business. Consider, for example, classification models in bioinformatics that classify catalytic protein structures as being in an active or inactive conformation. As an example from the field of medical informatics we might consider a classification model that, given the parameters of a tumor, will classify it as malignant or benign. Finally, a classification model in a bank might be used to tell the difference between a legal and a fraudulent transaction. Central to constructing, deploying, and using classification models is the question of model performance assessment (Hastie, Tibshirani, & Friedman, 2001). Traditionally this is accomplished by using metrics derived from the confusion matrix or contingency table. However, it has been recognized that (a) a scalar is a poor summary for the performance of a model in particular when deploying non-parametric models such as artificial neural networks or decision trees (Provost, Fawcett, & Kohavi, 1998) and (b) some performance metrics derived from the confusion matrix are sensitive to data anomalies such as class skew (Fawcett & Flach, 2005). Recently it has been observed that Receiver Operating Characteristic (ROC) curves visually convey the same information as the confusion matrix in a much more intuitive and robust fashion (Swets, Dawes, & Monahan, 2000). Here we take a look at model performance metrics derived from the confusion matrix. We highlight their shortcomings and illustrate how ROC curves can be deployed for model assessment in order to provide a much deeper and perhaps more intuitive analysis of the models. We also briefly address the problem of model selection.