Fatih Keskinkılıç

This study examines the temporal dynamics of environmental sustainability in G7 countries and evaluates how dominant temporal transformation regimes are associated with different environmental sustainability indicators. Existing studies on environmental sustainability in advanced economies have largely focused on individual indicators, particularly CO₂ emissions, or on average relationships estimated through conventional empirical models. However, sustainability transformation is multidimensional and may differ across production-based emissions, consumption-based ecological pressure, and ecosystem carrying capacity. To address this gap, this study analyzes annual data for G7 countries over the 1997–2021 period using three complementary indicators: CO₂ emissions, Ecological Footprint, and Load Capacity Factor. The study develops an integrated decision-support framework that combines MiniROCKET-based temporal feature extraction, Principal Component Analysis, and Double Machine Learning. First, macroeconomic, demographic, technological, energy-related, and environmental variables are transformed into high-dimensional temporal features using MiniROCKET. Second, PCA is used to reduce these features into dominant temporal components. Third, DML is applied to estimate the average and country-specific heterogeneous associations between the dominant temporal regime and the three environmental indicators. In the main specification, PC1 is used as the treatment variable, while PC2–PC5 are included as control components. Robustness checks include an alternative PC1–PC2 treatment definition, alternative nuisance learners, bootstrap confidence intervals, and placebo treatment tests. The PCA results show that the first five components explain approximately 89.92% of the total MiniROCKET-based temporal feature variance, indicating that G7 sustainability dynamics can be summarized by a limited number of dominant temporal structures. The main DML results indicate negative directional associations between the dominant temporal regime and all three environmental indicators. The strongest association is observed for CO₂ emissions, followed by Ecological Footprint and Load Capacity Factor. However, bootstrap confidence intervals include zero; therefore, the findings should be interpreted as indicative and decision-support-oriented evidence rather than definitive causal proof. Country-specific CATE results reveal substantial heterogeneity across G7 countries. Italy displays the strongest negative CATE values for both CO₂ emissions and Ecological Footprint, while the United Kingdom shows positive CATE values for these indicators. LCF effects remain close to zero across countries. These findings suggest that G7 sustainability policies should be country-specific and indicator-sensitive, integrating emission reduction, responsible consumption, circular economy practices, supply-chain transparency, and ecosystem-capacity-enhancing measures.

In casting processes, mechanical properties such as hardness are highly sensitive to both chemical composition and process parameters, making parameter design a complex and uncertain task during the embodiment stage of engineering design. Conventional trial-and-error-based approaches are often costly, time-consuming, and impractical in industrial environments. To address these challenges, this study proposes an optimized fuzzy artificial neural network (FANN)-based decision-support approach for hardness-oriented parameter design in a casting process. The developed model uses chemical composition variables, including carbon, silicon, manganese, phosphorus, sulfur, chromium, copper, and tin, together with process parameters such as casting temperature and casting time as inputs, while Brinell hardness is considered as the output. A dataset consisting of 170 experimental casting samples was employed; 128 samples were used for model development and hyperparameter selection, and 42 samples were reserved as an independent final test set. The proposed model was implemented as a scaled direct FANN weighted ensemble, in which fuzzified input variables were used to predict standardized continuous hardness values. A total of 300 FANN configurations were evaluated using five-fold cross-validation, and the five best-performing configurations were combined through RMSE-based weighted ensemble averaging. The final model was compared with Random Forest, Linear Regression, Ridge Regression, and SVR-RBF models using MSE, RMSE, MAE, R², MAPE, normalized RMSE, and ±5% prediction success rate. The results showed that the optimized FANN ensemble achieved the lowest mean RMSE in the full-data five-fold cross-validation analysis, slightly outperforming the Random Forest benchmark. In the independent final test set, Random Forest produced the lowest prediction error, whereas the proposed FANN ensemble remained competitive and achieved the same ±5% prediction success rate as Random Forest, Linear Regression, and Ridge Regression. Furthermore, a target-hardness case study demonstrated that the proposed approach could identify candidate casting conditions very close to a desired hardness level, with the nearest prediction reaching 202.985 HB for a target value of 203 HB. These findings indicate that the proposed FANN-based framework can serve not only as a hardness prediction model but also as a practical fuzzy decision-support tool for target-hardness-oriented parameter design in casting processes.