Integrating Multiple Techniques to Enhance Medical Data Classification

Abstract

Improving classification performance is an essential task in medical data classification. In the current medical data classification technique, if data pre-processing is not performed, the approach is more time consuming and has less classification accuracy. Here, the authors proposed two pre-processing techniques for enhancing the classification performance on medical data. The first pre-processing technique is noise filtering to improve the data quality. The second pre-processing bag of words technique is used for better feature selection. Subsequently, the hybrid fuzzy neural network approach is used for classification to handle data imprecision during classification. This arrangement of data pre-processing and the fuzzy neural classifier method improve classification accuracy.

Enhancing Prediction Accuracy in an Imbalanced Dataset of Dengue Infection Cases Using a Two-layer Ensemble Outlier Detection and Feature Selection

“…In most literature publications, constructing robust classification models involves improving data quality by mining excess outliers, resolving imbalanced instances of class data, and filtering out irrelevant features. The purpose is to enhance model training efficiency and prediction accuracy [4,8,9]. Due to the large dataset with numerous features, an outlier detection algorithm is required to minimize errors in checking and correcting the data.…”