Symptom-Based Disease Prediction: A Machine Learning Approach

Abstract

The advent of machine learning techniques has revolutionized various sectors, including healthcare. This project concentrates on leveraging machine learning algorithms for disease prediction based on symptoms. With a dataset comprising 132 symptoms and 41 diseases, the aim is to develop a robust predictive model capable of accurately diagnosing diseases given a set of symptoms. The process involves several key steps. Initially, the dataset is preprocessed to handle missing values, encode categorical variables, and normalize the data. To determine which symptoms are most pertinent to the prognosis of a disease, feature selection techniques are utilized. Various machine learning algorithms, including decision trees, support vector machines, random forests, and XGBoost, were explored to determine the most effective prediction model. XGBoost, in particular, emerges as one of the topperforming models because of its capacity to manage complicated relationships within the data and its effectiveness in handling imbalanced datasets. To evaluate the models' performance, evaluation criteria like accuracy, precision, recall, and F1-score are used. Moreover, to enhance model performance and avoid overfitting, techniques like crossvalidation and hyperparameter tuning are utilized. The proposed system holds significant potential in aiding healthcare professionals in diagnosing diseases promptly and accurately, thereby improving patient outcomes and reducing healthcare costs. It is important to note that the model needs further validation on diverse datasets and regular updates to remain relevant in clinical settings.