An Application of Using Support Vector Machine Based on Classification Technique for Predicting Medical Data Sets

Document Type

Conference Proceeding

Source of Publication

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Publication Date

1-1-2019

Abstract

© 2019, Springer Nature Switzerland AG. This paper illustrates the utilise of various kind of machine learning approaches based on support vector machines for classifying Sickle Cell Disease data set. It has demonstrated that support vector machines generate an essential enhancement when applied for the pre-processing of clinical time-series data set. In this aspect, the objective of this study is to present discoveries for a number of classes of approaches for therapeutically associated problems in the purpose of acquiring high accuracy and performance. The primary case in this study includes classifying the dosage necessary for each patient individually. We applied a number of support vector machines to examine sickle cell data set based on the performance evaluation metrics. The result collected from a number of models have indicated that, support vector Classifier demonstrated inferior outcomes in comparison to Radial Basis Support Vector Classifier. For our Sickle cell data sets, it was found that the Parzen Kernel Support Vector Classifier produced the highest levels of performance and accuracy during training procedure accuracy 0.89733, AUC 0.94267. Where the testing set process, accuracy 0.81778, the area under the curve with 0.86556.

ISBN

9783030269685

ISSN

0302-9743

Publisher

Springer Verlag

Volume

11644 LNCS

First Page

580

Last Page

591

Disciplines

Computer Sciences

Keywords

Classification, Evaluation techniques, Machine learning, Sickle cell disorder data set, Support vector machines

Scopus ID

85070554029

Indexed in Scopus

yes

Open Access

no

Link to Full Text

Share

COinS