HEAD INJURY RISK PREDICTION USING MACHINE LEARNING

Authors

  • POREDDY JAYARAJU1 , VENKATA SNEHITH2 , SAI KUMAR3 , TEJESHWAR SAI4 , ABHILASH5 Author

DOI:

https://doi.org/10.5281/zenodo.21369315

Abstract

Head injury is one of the leading causes of severe trauma and mortality worldwide, making early and accurate risk assessment essential for effective medical treatment. Traditional methods of identifying the severity of head injuries rely on clinical observation and manual evaluation of multiple patient parameters, which can be time-consuming and may lead to delayed decision-making in emergency situations. To address this challenge, the proposed project, "Head Injury Risk Prediction Using Machine Learning," presents an intelligent decisionsupport system that predicts the risk level of head injury using a supervised machine learning approach. The proposed system employs the Logistic Regression algorithm, a simple, interpretable, and computationally efficient supervised learning model, to classify patients into High Risk or Low Risk categories. The prediction is based on important clinical parameters collected during the initial medical assessment, such as patient age, Glasgow Coma Scale (GCS) score, loss of consciousness, vomiting, skull fractures, amnesia, and other relevant indicators associated with traumatic brain injury. These parameters are provided through a user-friendly web-based application developed using Flask, HTML, and CSS, allowing healthcare professionals to enter patient information quickly and receive immediate predictions. The system preprocesses the input data, applies the trained Logistic Regression model, and generates accurate risk predictions within a short time. Compared to traditional manual assessment methods, the proposed solution reduces decision-making time, minimizes human error, and provides consistent and reliable predictions that assist medical professionals during the initial triage process. The lightweight nature of the Logistic Regression algorithm also makes the system suitable for deployment on resourceconstrained environments without requiring high computational power. The performance of the proposed model is evaluated using standard machine learning evaluation metrics such as accuracy, precision, recall, F1-score, and confusion matrix to ensure reliable classification performance. The results demonstrate that Logistic Regression provides an effective balance between prediction accuracy, interpretability, and computational efficiency, making it an appropriate choice for healthcare decisionsupport applications. The proposed system has the potential to improve early clinical decision-making, support emergency medical personnel in prioritizing patient care, and contribute to faster diagnosis and treatment of traumatic brain injuries.

Downloads

Published

15-07-2026

How to Cite

HEAD INJURY RISK PREDICTION USING MACHINE LEARNING. (2026). International Journal of Engineering Research and Science & Technology, 22(2(4), 1764-1773. https://doi.org/10.5281/zenodo.21369315