Ant Colony Optimization Driven Failure Prediction in Industrial Internet of Things

In today's Industrial Internet of Things (IIoT), need arises for accurate Failure Prediction (FP). The research work proposes the adaptation of a nature-inspired optimization algorithm in the IIoT application for FP. This research work is focused on the application of the Ant Colony Optimization (ACO) algorithm to enhance features selection. For this research work, AI4I Predictive Maintenance Dataset is used. Support Vector Machines (SVM), Decision Trees, K-Nearest Neighbors (KNN), Logistic Regression (LR), and Neural Networks (NN) these five models are trained using the features retrieved by ACO.So these models are called ACO driven models. Two step evaluation process is used. First is without hyperparameter tuning and second with hyperparameter tuning. Comparison of obtained results with existing work is done. Accuracy ranging from 96.75% to 99.90% is obtained by ACO driven models, such as SVM, DT, KNN, LR, and NN. Superior performance of proposed ACO driven models has provided valuable insights into advancing FP for IIoT systems.