Dengue fever prediction using machine learning techniques

Abstract

The rapid and broad spread of the virus that causes dengue is a hallmark of the dengue, a global epidemic, a worldwide health problem. Using data from Kaggle, this project is about predicting Dengue Fever through machine learning approaches. Dataset has 33 variables, contain a variety of symptoms like itching, rashes on the skin, joint discomfort, high temperature, and more. "Prognosis," the target attribute, divides occurrences into three classes: typhoid, common cold, and dengue, with counts of 209, 143, and 103, respectively. AdaBoostClassifier, BernoulliNB, GaussianNB, DecisionTreeClassifier, BaggingClassifier, as well as Voting Classifier are a few of the machine learning models used for prediction. The BaggingClassifier algorithm performed quite well in this instance, with the greatest accuracy of 95.87%. The process is methodical and starts with the selection of information from Kaggle. Next, preprocessing operations such as encoding and missing value removal are carried out. EDA, or exploratory data analysis, provides light on the properties of the dataset. Algorithms such as AdaBoostClassifier, DecisionTreeClassifier, and many more are used in model training. Evaluation measures measure the performance of the models and include recall, accuracy, and precision. Testing the algorithms to determine how well they can forecast the real world is the last phase