Resti Amalia, Ahmad Faiz Zaidan, Syahrul Ramadhan, Farhan Septian, Ananta Mikail Aqsha, Perani Rosyani

Classification of Autoimmune Diseases Using the K-Nearest Neighbors Algorithm

Introduction

Classification of autoimmune diseases using the k-nearest neighbors algorithm. Classify autoimmune diseases using the K-Nearest Neighbors (K-NN) algorithm. This study reviews K-NN's high accuracy, challenges, and enhanced effectiveness with hybrid optimization.

Abstract

Autoimmune diseases occur when the immune system attacks the body’s own tissues, causing serious complications and overlapping symptoms that challenge early detection. This study reviews the use of the K-Nearest Neighbors (K-NN) algorithm for classifying autoimmune diseases through a systematic literature review of five articles. Compared to methods like Genetic Algorithms, Support Vector Machines (SVM), and Single Layer Perceptrons (SLP), K-NN shows high accuracy when optimal parameters and neighbor counts are used. However, challenges include sensitivity to imbalanced data and high computational demands for large datasets. Combining K-NN with optimization techniques, such as Genetic Algorithms, enhances accuracy and stability. The study concludes that K-NN is effective for classifying autoimmune diseases, especially with hybrid approaches, and recommends further research with larger datasets.

Review

This systematic review addresses a highly relevant and challenging area: the classification of autoimmune diseases, which are notoriously difficult to diagnose due to overlapping symptoms and the complexity of the immune system. The decision to focus on the K-Nearest Neighbors (K-NN) algorithm is appropriate, given its widespread use in machine learning. The abstract effectively highlights K-NN's reported strengths, such as high accuracy when optimal parameters are employed, and its competitive performance against other algorithms like SVM and SLP within the scope of the reviewed literature. The acknowledgment that hybrid approaches can further enhance accuracy and stability is a valuable insight, suggesting a promising direction for practical application. However, the methodology described raises significant concerns regarding the scope and generalizability of the findings. A "systematic literature review of five articles" is an exceptionally narrow foundation for drawing robust conclusions. This limited number of studies makes it difficult to ascertain the breadth of evidence, potential biases in the literature, or the consistency of K-NN's performance across diverse autoimmune diseases and datasets. While the abstract correctly identifies inherent challenges of K-NN, such as sensitivity to imbalanced data and high computational demands for large datasets, these are well-known limitations of the algorithm itself, rather than novel findings from a comprehensive review. Greater detail on the selection criteria for these five articles, their specific characteristics, and how their findings were synthesized would be crucial to assess the rigor of this review. Despite these methodological limitations, the study concludes by reaffirming K-NN's effectiveness for classifying autoimmune diseases, particularly when integrated into hybrid models. The recommendation for further research with larger datasets is critical, not only for empirical studies but also for future systematic reviews to provide a more comprehensive and statistically robust understanding of K-NN's utility in this domain. This review serves as a preliminary exploration of K-NN's application in autoimmune disease classification, pointing towards its potential while underscoring the necessity for a much broader evidence base and more rigorous methodological approaches in future research to validate and expand upon these initial findings.

Full Text

Comments

You need to be logged in to post a comment.