Gold is an important investment instrument of choice around the world. At the same time, its value and growth potential are of great interest. The appeal of this precious metal lies in its intrinsic value and potential for lucrative growth over time, making it a subject of intense interest and analysis among investors and economists alike. One of the key determinants of economic and social dynamics is the volatility of exchange rates and the gold market. Therefore, understanding and correctly interpreting the fluctuations of exchange rates and the gold market is of great importance for both individual investors and corporate firms. In this study, the factors affecting gold prices were investigated and the correlation between the values found and gold prices was revealed, and it was aimed to predict the change in gold price in the future with Artificial Neural Networks (ANN). The data set used in the study consists of data between January 2, 2020, and March 15, 2024. Since the markets were closed on holidays and weekends, these dates were separated and a total of 1005 days of data were used. The developed model was created and tested with different layers and neurons using MATLAB Neural Net Fitting. Results were evaluated using Coefficient of Determination (R²), Mean Squared Error (MSE) and Mean Absolute Error (MAE). The best success rate was obtained by using Levenberg-Marquardt algorithm, hyperbolic tangent sigmoid activation function in hidden layers and “purelin” linear activation function in the output layer, 4 layers and 3 neurons and 94.06% success rate was observed.
This research scrutinizes the abrupt behavioral shifts surrounding digital financial platforms in the Turkish sector, utilizing a multidimensional forecasting apparatus that fuses classical regression, seasonal econometrics, and tree-based predictive learners. Fifty-three quarters of aggregate usage logs (late-2011 to late-2024) were processed via a tiered methodology: foundational polynomial estimators mapped the continuous growth baseline, while complex seasonal autoregressive models (SARIMA) absorbed cyclic temporal echoes. To address erratic nonlinear surges, ensemble protocols—specifically Random Forest and XGBoost—were subsequently deployed alongside an innovative composite framework that mitigates structured econometric residuals via iterative gradient boosting. Our quantitative results isolate a violent, statistically verified inflection point triggered during the second quarter of 2020, an anomaly that persistently overwhelmed standard linear projection boundaries. Comparatively, the integrated SARIMA–XGBoost construct achieved unmatched tracking accuracy (RMSE: 214,586; MAE: 122,847; MAPE: 2.6%; 𝑅2: 0.995), dramatically eclipsing the precision of isolated ARIMA frameworks. Ultimately, this analysis corroborates the assertion that melding standard time-series inference with algorithmic error-correction produces extraordinarily resilient predictive tools, particularly when modeling markets destabilized by sudden macroeconomic shocks.
The Acid Orange 7 (AO7) azo dye poses a serious problem due to its low environmental persistence and low treatment efficiency. Electrooxidation (EO) is a sustainable and highly efficient treatment technology for the effective oxidation of toxic AO7 azo dye. In this study, we investigated key parameters of the EO process for the oxidation of the AO7 azo dye, including current density (j), initial pH, initial AO7 concentration (Ci), and supporting electrolyte concentration (SEc). Under optimal operating conditions (anode/cathode = Pt/Gr-PtO2, j = 100 A/m2, pH = 3, Ci = 35 mg/L, and Na2SO4 = 50 mM), a removal efficiency of 79.0 % was achieved. Under the same conditions, the energy consumption and operating costs were calculated as 3.21 kWh/m3 (1471.82 kWh/kg AO7) and 0.57 $/m3 (11.96 $/kg AO7), respectively. Furthermore, the first-order model was evidenced by the determination of correlation coefficient values of (R2 = 0.984, AdjustedR2 = 0.963) and rate constant k = 0.00812 𝑚𝑖𝑛−1. These findings demonstrate that the EO process offers an effective and economical solution for removing AO7 azo dye, thereby supporting the potential of advanced oxidation technologies in environmental engineering.
Assembly line balancing models aim at minimizing cycle time, and the number of stations becomes quite complex with precedence relations and various resource constraints. Various heuristic and fuzzy-based methods are used to ease the solution in such assembly line balancing models. This study proposes a two-stage model that simultaneously minimizes cycle time and the number of stations for a real-life case. The paper shows how to formulate a mixed integer non-linear programming model for an assembly line balancing problem. After the constraints on production size and task durations are fuzzy transformed, the model acquires new constraints. In the new model, the α coefficient takes values in the range [0,1], bringing the fuzzy parameters closer to the lower or upper range. In cases where uncertain task duration and production size parameters are fuzzy, better results are achieved with the fuzzy model than the basic model. The fuzzy model provides to expand value range for parameters and evaluates different solution alternatives. In the study, the reflection of the value change in the parameters on the solution was examined with the experimental data set created in line with 112 different α coefficients. One-way ANOVA test was used to understand whether the fuzzy parameter values created a significant difference in the solution. The proposed model improves the solution quality through scenarios, provides a 7% saving for the total duration of the tasks at the stations within a cycle time and suggests a more suitable station layout for efficient production.
In this study, the non-isothermal decomposition kinetics of acrylonitrile–butadiene–styrene (ABS) and flame-retardant-modified ABS composites were investigated. Neat ABS, red phosphorus (RP)-containing, and RP/zinc borate (ZnB)-containing composites were prepared by melt blending. Thermogravimetric analysis was performed at different heating rates under an inert atmosphere to evaluate the thermal degradation characteristics of the materials. The thermal decomposition kinetics were analyzed using the model-free Kissinger–Akahira–Sunose (KAS) and Flynn–Wall–Ozawa (FWO) isoconversional methods. The results indicate that the incorporation of flame-retardant additives significantly alters the thermal decomposition behavior of ABS. While neat ABS exhibits a relatively simple degradation process, the presence of RP leads to a more complex decomposition mechanism, particularly at early stages, due to the formation of a protective char layer. The hybrid RP/ZnB system exhibits a moderate kinetic response, suggesting a synergistic effect that stabilizes the degradation process over a broader conversion range. Overall, the study demonstrates that non-isothermal kinetic analysis is a practical approach for evaluating the influence of flame-retardant additives on the thermal degradation behavior of ABS composites and provides valuable insights into the thermal degradation behavior of polymer systems containing RP and ZnB additives.
