Support vector machine (SVM) algorithms have been widely used for classification in many different areas. However, the use of a single SVM classifier is limited by the advantages and disadvantages of the algorithm. This paper proposes a novel method, called support vector machine chains (SVMC), which involves chaining together multiple SVM classifiers in a special structure, such that each learner is constructed by decrementing one feature at each stage. This paper also proposes a new voting mechanism, called tournament voting, in which the outputs of classifiers compete in groups, the common result in each group gradually moves to the next round, and, at the last round, the winning class label is assigned as the final prediction. Experiments were conducted on 14 real-world benchmark datasets. The experimental results showed that SVMC (88.11%) achieved higher accuracy than SVM (86.71%) on average thanks to the feature selection, sampling, and chain structure combined with multiple models. Furthermore, the proposed tournament voting demonstrated higher performance than the standard majority voting in terms of accuracy. The results also showed that the proposed SVMC method outperformed the state-of-the-art methods with a 6.88% improvement in average accuracy.

Machine learning and data mining techniques provide businesses with cutting-edge data-driven decision-making capabilities. Their popularity is growing because they enable more accurate and consistent evaluation and prediction of current and future situations based on previous data. This study used machine learning methodologies to address three of the textile industry's most pressing concerns: lead time, cloth waste, and price. A multi-output regressor model in which three subjects are predicted simultaneously is also investigated, in addition to training individual models for each subject. XGBoost is the model with the best lead time prediction results, with an R2 of 0.86 and an MAE of 8.35. When all three subjects are predicted at the same time, XGBoost achieves R2 of 0.88 and MAE of 3.79. These findings indicate that, in addition to single models, multi-output models are also promising.

The aim of this study was to develop a method to dye polyethylene terephthalate fabrics to light and medium shades at 95 °C without the use of carriers, dispersing agents, or reducing agents and to evaluate the effectiveness of monocationic and dicationic ionic liquids as auxiliary chemicals in the dyebath. Four ionic liquids were employed: 1-butyl-3-methylimidazolium chloride, 1-butyl-3-methylimidazolium bromide, 1,4-bis(3-methylimidazolium)butane dibromide, and 1,4-bis(3-methyl imidazolium)butane dichloride. Dye exhaustion increased from 31.5% in the IL-free control sample to 88.4% in the presence of 10 g/L of the dibromide-based dicationic ionic liquid. Although different crystallinity values were obtained using different analytical techniques, both methods indicated lower crystallinity in the corresponding dyed sample than in the control. Considering the nonlinear nature of polyethylene terephthalate fabric dyeing, machine learning models were employed to predict color strength based on ionic liquid type, concentration, and measured color coordinates. The models’ performance was evaluated using five-fold cross-validation and complementary statistical metrics. Shapley Additive Explanations analysis was employed to interpret the contributions of input variables to model predictions. The results identified ionic liquid concentration and type as the most influential features in the selected model, followed by colorimetric descriptors.