International Journal of Research and Innovation in Applied Science (IJRIAS) |Volume VII, Issue X, October 2022|ISSN 2454-6194
Estimating The Accuracy of Classifiers in Analyzing Multiple Diseases
AJAYI Olusola Olajide
Department of Computer Science, Faculty of Science, Adekunle Ajasin University, Akungba-Akoko, Ondo, Nigeria
Abstract: Medical data are regarded as been sensitive not only in terms of the need to keep it private but also and majorly in terms of the need to get it right and accurate. Patients’ medical data are diagnose and analyze with optimal accuracy to avoid error of prescription. Multiple diseases are one that can easily get complicated where the analysis of symptoms are not right. Machine learning is a known field of inquiry found very suitable in the medical area for analysis of medical diagnosis. The need for the right classification algorithm to deploy for a particular medical experimentation/prediction becomes very germane especially in the case of multiple diseases. No doubt, many researches have been done in this regard but not specifically tailored towards multiple diseases. The study which utilizes medical data from third party, www.kaggle.com, applied selected common three classification algorithms on the dataset. The result of the experimentation carried out using WEKA Explorer, shows Artificial Neural Network (ANN) outperforms Decision Tree and Naïve Bayes in terms of level of accuracy.
Keywords: Classifier, Algorithms, Accuracy, Multiple Diseases, Morbidity, Prediction
I. INTRODUCTION
Diseases are generally understood to be medical conditions that involve a pathological process associated with a specific set of symptoms. Localized diseases affect specific parts of the body; disseminated diseases spread to other parts of the body; and systemic diseases affect the entire body. Each disease process has an origin, or etiology, but some diseases may present with different or confusing symptoms, making them difficult to diagnose or determine. The physical symptoms of disease may be accompanied by emotional symptoms, and some diseases that affect the chemical balances of the nervous system may manifest in physical symptoms.
Disease prediction mean some numbers of symptoms are selected for processing and using this symptom as input we can predict the disease with help of many kind of classification algorithms. Several data mining algorithms and techniques have been used for study and analysis of various diseases like Hepatitis, Diabetes and Cancer etc. Recent survey shows that heart disease is one of the biggest causes of death in the countries like UK, Canada, France and Singapore. Various classification models like Decision Tree, KNN and Naive Bayes have been used to diagnose the presence of diseases in patients. Several researches have been carried out in the area of disease prediction among which include An Ensemble Multilabel Classification for Disease Risk Prediction (Runzhi et al., 2017), Predicting disease risks from