Alzheimer’s disease prediction using three machine learning methods

Abstract

Alzheimer's disease (AD) is the most common incurable neurodegenerative illness, a term that encompasses memory loss as well as other cognitive abilities. The purpose of the study is using precise early-stage gene expression data from blood generated from a clinical Alzheimer's dataset, the goal was to construct a classification model that might predict the early stages of Alzheimer's disease. Using information gain (IG), a selection of characteristics was chosen to provide substantial information for distinguishing between normal control (NC) and early-stage AD participants. The data was divided into various sizes; three distinct machine learning (ML) algorithms were used to generate the classification models: support vector machine (SVM), Naïve Bayes (NB), and k-nearest neighbors (K-NN). Using the WEKA software tool and a variety of model performance measures, the capacity of the algorithms to effectively predict cognitive impairment status was compared and tested. The current findings reveal that an SVM-based classification model can accurately differentiate cognitively impaired Alzheimer's patients from normal healthy people with 96.6% accuracy. As discovered and validated a gene expression pattern in the blood that accurately distinguishes Alzheimer's patients and cognitively healthy controls, demonstrating that changes specific to AD can be detected far from the disease's core site.

Feature selection optimization based on genetic algorithm for support vector classification varieties of raisin

“…There are 3 algorithms used, namely SVM, Naïve Bayes (NB), and K-nearest neighbors (K-NN). The current findings reveal that the SVM-based classification model can accurately distinguish cognitively impaired Alzheimer's patients from normal healthy individuals with 96.6% accuracy [30]. In this study, the classification of the Besni and Kecimen raisin varieties produced in Turkey was carried out using the SVM algorithm with a dataset of 900 data.…”

A Machine Learning Model for Cancer Disease Diagnosis using Gene Expression Data

“…Microarray gene expression data processing is one of the difficult study subjects in the fields of Computational Biology, Genomics, Statistics, and Pattern Classification. The main problem with microarray cancer analysis is the short sample size and high curse of dimensionality brought on by redundant and irrelevant genes [3] [4] . Additionally, the majority of medical datasets exhibit noise, varying feature values, and an unbalanced number of classes, all of which contribute to over-fitting and reduced classification accuracy [5] [6].…”

Optimizing blockchain technology using a data sharing model

“…While the development of big data technologies has made life more convenient for people, various data security issues have also been brought up [8], [9]. The constant occurrence of security events like data management security borders and resource theft is one of the most significant issues [10], [11]. Resource hurdles that have been created as a result of data security, and realizing the value of data can be challenging due to business considerations and privacy issues, additionally, the financial benefits of comparing various forms of data do not fully materialize.…”