A comparative analysis of machine learning methods for classification type decision problems in healthcare

Abstract

Advanced analytical techniques are gaining popularity in addressing complex classification type decision problems in many fields including healthcare and medicine. In this exemplary study, using digitized signal data, we developed predictive models employing three machine learning methods to diagnose an asthma patient based solely on the sounds acquired from the chest of the patient in a clinical laboratory. Although, the performances varied slightly, ensemble models (i.e., Random Forest and AdaBoost combined with Random Forest) achieved about 90% accuracy on predicting asthma patients, compared to artificial neural networks models that achieved about 80% predictive accuracy. Our results show that noninvasive, computerized lung sound analysis that rely on low-cost microphones and an embedded real-time microprocessor system would help physicians to make faster and better diagnostic decisions, especially in situations where x-ray and CT-scans are not reachable or not available. This study is a testament to the improving capabilities of analytic techniques in support of better decision making, especially in situations constraint by limited resources.

Two-level coarse-to-fine classification algorithm for asthma wheezing recognition in children's respiratory sounds

“…A similar accuracy has been obtained using advanced signal processing techniques based on the persistent homology of delay embedding [3]. Although the literature mentions the importance of ensemble classifiers in the decision making related to asthma patients [18,19], there is no paper, based on our knowledge, that investigates more deeply the efficiency of ensemble methods in wheezing detection in comparison to individual classifiers. In addition, the majority of the experimental results are evaluated based on the breathing record acquired under the controlled condition from the adult patients [29].…”

Assessing Advanced Machine Learning Techniques for Predicting Hospital Readmission

“…Furthermore, it is simple, swift, and easy to use with an iterative algorithm which requires only one parameter, iteration number. Moreover, it is not subject to over-fitting and simply determines outliners which are incorrectly classified or are difficult to classify [21]. However, misclassified and/or difficult instances are given significance by gradient boosting (GB), via the remaining errors-also known as pseudo-residuals-of a strong learner.…”

Class imbalance in gradient boosting classification algorithms: Application to experimental stroke data

“…Both strategies have limitations such as loss of data with undersampling or overfitting with oversampling, and the effect of resampling has rarely been evaluated. 9 Synthetic Minority Oversampling Technique (SMOTE) is one way of achieving class balance. It is designed to generate new synthetic data that is coherent with the minority class distribution while minimising overfitting.…”