CNOL: A Convoluted Neural Optimization Logic to Predict Leptospirosis Disease Detection based on Medical Data Evaluation Scheme

Abstract

Leptospirosis is a widespread zoonotic disease that can cause severe health issues if not detected and treated early. This study presents a Convoluted Neural Optimization Logic (CNOL) framework designed to predict Leptospirosis Disease using both structured and unstructured medical data. The proposed model integrates convolutional layers for feature extraction from complex datasets, including clinical information and medical images, and utilizes advanced optimization techniques like Adam to enhance learning. The CNOL framework tested on a Leptospirosis Disease Dataset, and results show a significant improvement in predictive accuracy. The model achieved an accuracy of 95.67%, outperforming traditional machine learning models such as Logistic Regression (85.45%) and Random Forest (89.56%), and deep learning models like Deep Neural Networks (93.10%). Additionally, the CNOL framework exhibited superior precision (94.85%) and recall (96.42%), highlighting its effectiveness in both identifying true positive cases and minimizing false negatives. This study demonstrates the potential of combining convolutional layers and advanced optimization techniques for enhancing disease prediction, providing a robust tool for early diagnosis and intervention.