Detection of meningitis disease using Belief Bidirectional Neural Network and Informative Ant Colony Optimization techniques

Meningitis is a serious illness brought on by inflammation of the membranes that edge the brain and spinal cord. To lower the risk of serious complications and death early and precise diagnosis is crucial especially for bacterial meningitis. Conventional diagnostic methods on the other hand frequently have poor accuracy lag in processing and a failure to evaluate the marginal influence of disease characteristics. This study suggests a novel hybrid framework that combines Informative Ant Colony Optimization (IACO) and Belief Bidirectional Neural Network (B2N2) for efficient meningitis detection in order to overcome these limitations. The first step in the suggested system is data preprocessing which uses Z-Score Normalization (ZSN) to scale the dataset and eliminate outliers. The marginal contribution of each feature is then estimated using the Meninges Affect Rate (MAR) algorithm. The IACO approach optimizes feature selection to improve classification relevance based on MAR scores. Lastly the B2N2 model uses a belief-driven bidirectional learning approach to classify the data. The suggested framework outperforms current techniques like SegResNet Gradient Boosted Trees (GBT) and Multiple Logistic Regression (MLR) with an improved classification accuracy of 94–25% according to experimental results. The framework also performs better on important metrics like time complexity F1- score recall and precision. These outcomes demonstrate the B2N2-IACO approaches potential as a scalable and trustworthy diagnostic method for meningitis detection in real time.