Evaluation of the Performance of Machine Learning Algorithms in Disease Prediction

Abstract

Today, machine learning is widely applied in various disciplines such as technology, healthcare, law, cybersecurity, and image recognition. When examining the research, it is evident that the scope of machine learning applications is expanding day by day. In this study, the goal was to develop a classifier model using machine learning algorithms for disease diagnosis in the healthcare field. In the scope of the study, the performance of various machine learning algorithms such as Naive Bayes, Support Vector Machines (SVM), Decision Trees (CART), Random Forest, Gradient Boosting, and AdaBoost was compared for disease prediction. The dataset used in the study was obtained from the Kaggle platform and includes records where diseases are predicted based on various symptoms.The dataset is organized into two different CSV formats for training and testing. The training dataset was used for the model’s learning process, while the testing dataset was used to evaluate the accuracy and performance of the model. The dataset contains a total of 4,962 records and consists of 133 columns, with 132 independent variables (symptoms) and 1 dependent variable (disease) for classification. The dataset includes 41 different diseases, and there are 120 examples for each disease. When comparing the accuracy performance of the algorithms used in the study, the highest success rates were achieved with Naive Bayes, Support Vector Machines (SVM), and Gradient Boosting algorithms. Jupyter Notebook was used in the processes of data preparation and model development.