Detection of Petrochemical Pollutants using Multi Mutation Tunicate Swarm Optimization Algorithm based Convolutional Neural Network with Leaky Rectified Linear Unit

Environmental contamination does not prevent, and one such source is hazardous petroleum waste. These pollutants' leachates pollute drinking water supplies and soil, endangering public health. In this research aims to develop the Multi mutation-based Tunicate Swarm Optimization (MMTSO) based Convolutional Neural Network (CNN) with Leaky Rectified Linear Unit (LReLU) for early prediction of the risk factors of petroleum content the proposed model names as a MMTSO-CNN-LReLU. Initially the dataset collected from the existing paper and it’s taken to the preprocessing stage for remove the unwanted noise by using normalization and noise removal technique. The Principal Component Analysis (PCA) is utilized to extract features and their store in a vector, then the proposed method of multi-mutation based Tunicate Swarm Optimization Algorithm is used to select the relevant features. The proposed MMTSO is effectively contributing feature engineering part and other hand which is improves the performance of the CNN classifier. Finally, the selected features are fed to the CNN with LReLU model to predict the rick factor of Petrochemical content. The proposed MMTSO method shows the accuracy of 98.47%, precision of 95.78% and recall of 97.58% compared to existing methods.