Multilevel Weighted Support Vector Machine for Classification on Healthcare Data with Missing Values

Abstract

This work is motivated by the needs of predictive analytics on healthcare data as represented by Electronic Medical Records. Such data is invariably problematic: noisy, with missing entries, with imbalance in classes of interests, leading to serious bias in predictive modeling. Since standard data mining methods often produce poor performance measures, we argue for development of specialized techniques of data-preprocessing and classification. In this paper, we propose a new method to simultaneously classify large datasets and reduce the effects of missing values. It is based on a multilevel framework of the cost-sensitive SVM and the expected maximization imputation method for missing values, which relies on iterated regression analyses. We compare classification results of multilevel SVM-based algorithms on public benchmark datasets with imbalanced classes and missing values as well as real data in health applications, and show that our multilevel SVM-based method produces fast, and more accurate and robust classification results.

Application of support vector machine modeling and graph theory metrics for disease classification

“…SVM has been used across diverse biomedical classification problems. This includes a patient financial risk model using health claims and clinical encounter data, and a patient response to flu awareness campaign model, both using weighted SVM [10]. A project comparing various machine learning techniques with Logistic Regression for prediction of heart disease also shows no significant difference between Logistic Regression and SVM, with the Linear kernel performing best [11].…”

Predictive models for bariatric surgery risks with imbalanced medical datasets

“…Our work lies in the medical pattern recognition framework, which is known to be highly imbalanced, i.e., the instances of interest in the dataset are relatively rare. We discuss the hardness of classification problems with healthcare imbalanced data with missing values in [44]. Examples are Intense Care Unit (ICU) infection detection events [46], medical diagnoses [34], adverse drug events [51,48], bleeding detection in endoscopic video [17], and so on.…”

What does the mind learn? A comparison of human and machine learning representations

“…One simple form this method can take is to store all experienced items in a representational space with a similarity gradient around each for use in future prediction or classification. This is exemplified in recent kernel methods, which use a variety of similarity metrics and often remove redundant stored items to provide a more efficient representation [1]; for example, support vector machines use kernel functions to draw decision boundaries between categories [2], often resulting in highly accurate classification performance [3][4][5].…”