Forecasting Future Trajectories in Rubber Technology Using Machine Learning

In the abstract, To achieve optimal performance and reliability, rubber product manufacturers must implement severe quality control techniques. In this paper, researchers suggest and develop a unique defect detection methodology for rubber technology that uses a combination of CNN (Convolutional Neural Networks) and Random Forest algorithms. The research focuses on six unique defect types: tear-resistant deficiencies oxidation damage, pitting, hardness spots, flashing, or fisheyes/pinholes. After thorough experimentation with evaluation, our model performs exceptionally well, with accuracy ratings that vary from 95.29% to 96.22%, recall values ranging from 95.24% to 96.05%, and Fl scores ranging from 95.52% to 95.96%. These measures, together with the high support values, demonstrate the model's accuracy or robustness in defect diagnosis. Furthermore, supporting proportion analysis provides information about the distribution of cases across defect types. The total accuracy of 99% demonstrates the model's ability to correctly understand rubber problems. it suggested approach not only improves fault identification in rubber technology, but it also offers a prospective path for developing quality assurance operations, resulting in increased product reliability and lower manufacturing costs.