Fuzzy Machine Learning Model in Real-World Physical Domains; A State-of-the-Art Approach

Day after day, individuals all across the world come up with fresh ideas for improving the future. Several intriguing discoveries and ideas paved the way for a new age of electronics, telecommunications, business, and medicinal innovation. Using less resources, greater changes in these domains can be achieved. As improving efficiency and productivity allow exponential development in some areas of the global economy, artificial intelligence (AI) and machine learning (ML) is being adopted by a growing number of individuals, corporations, and governments. Since real-world scenarios influence imprecise and unpredictable situations, fuzzy systems have become an inescapable machine learning aspect. Thus, this research presents a qualitative analysis of the significance of fuzzy machine learning systems like fuzzy support vector machines (FSVM) in various physical domains. Based on this analysis, this research extends with the proposal of a fuzzy machine learning-based framework for two different physical domains: (1) intelligent transportation and (2) ecological risk handling. Thereby, this state-of-the-art approach presents fault detection using FSVM (FD-FSVM) model in the intelligent transportation domain. In ecological risk handling, this study proposes an improved FSVM for risk level classification (FSVM-RLC) approach, which uses the persistent organic pollutants data for training and validation. These two domains are chosen randomly to evaluate the classification performance of the fuzzy machine learning algorithms based on their mean absolute error, accuracy, precision, recall, and F1 score. Apart from this, the mean square error and mean absolute error are measured. Compared to existing machine learning models, the individual results of these two approaches show the highest performance. Furthermore, this fuzzy integrated machine learning technique kept consistency in both domains by giving 98.2% and 97.89% accuracy levels for FD-FSVM and FSVM-RLC, respectively.