Urban perception is a multidimensional phenomenon reflecting individuals’ evaluations of the urban environment and playing a critical role in planning and design processes aimed at improving quality of life. This study aims to predict six different themes of urban perception (beautiful, boring, depressing, lively, safe, wealthy) from street view images using regression-based deep learning methods. Three different deep learning architectures—ResNet18, VGG19, and EfficientNet-B1—were employed. The Place Pulse 2.0 dataset was utilized in the modeling process, with approximately 110,000 labeled street images processed through necessary preprocessing steps (resizing, cropping, tensor conversion, and normalization). Models were trained with an 80% training and 20% validation split. Performance evaluation was conducted using Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), R2 and validation loss graphs. Findings indicate that the EfficientNet-B1 model achieved the lowest error values, particularly in the “safe” and “lively” themes, while the ResNet18 model offered more balanced and stable performance in terms of validation loss. The VGG19 model generally yielded higher error rates and exhibited a clear tendency toward overfitting. It was observed that theme-specific visual complexity directly affected model performance. In conclusion, while deep learning architectures prove effective in modeling urban perception through visual data, both the choice of architecture and the inherent nature of the theme play decisive roles in model performance. This study highlights the importance of architecture- and theme-sensitive model design in AI-supported analysis of urban perception.

This study investigates the differential geometric properties of Viviani's curve using the Darboux frame apparatus. Viviani's curve, a classical space curve arising from the intersection of specific surfaces, is examined from two distinct geometric perspectives: first as the intersection of a sphere and a circular cylinder, and second as the intersection of a circular cone and a parabolic cylinder. For each representation, the Darboux frame field consisting of the tangent vector, surface normal, and their cross product is explicitly constructed. The geodesic curvature, normal curvature, and geodesic torsion are derived and analyzed in detail. It is proven that Viviani's curve becomes a geodesic on the circular cylinder at specific parameter values (s=2kπ,k∈Z), while on the circular cone, the curve exhibits asymptotic behavior at s=kπ/2 and principal curve characteristics at s=kπ. The relationship between Darboux curvatures and the Frenet curvature is established, providing an alternative computational approach to classical Frenet-Serret formulas. Several illustrative examples demonstrate the Frenet and Darboux frames at specific points on the curve, revealing geometric insights about frame coincidence and orthogonality properties. Additionally, a double helix-like structure is constructed using two Viviani curves. This work contributes to the geometric understanding of Viviani's curve through the lens of surface-curve interaction theory and extends the theoretical framework for analyzing curves lying on classical surfaces.

Deep learning has emerged as a widely applied approach across various fields, with finance and forecasting being among its most prominent areas of use. Within this domain, different deep learning architectures have been developed to address specific prediction problems.This study compares the performance of ARIMAX and several deep learning models—including LSTM, BILSTM, CNN-LSTM, GRU, and TFT—in forecasting Bitcoin prices. The dataset consists of daily values from January 2014 to January 2025. The dependent variable is the daily Bitcoin closing price ($), while the independent variables include oil price (USD/barrel), gold price (USD/ounce), platinum price ($/XPT), and the USD/TRY exchange rate.All analyses were conducted in Python using Google Colab, with the Keras library employed for model implementation. Root Mean Square Error (RMSE) was selected as the evaluation metric for predictive accuracy.The results indicate that the TFT model achieved the highest predictive performance, followed closely by the GRU model. LSTM, BILSTM, and ARIMAX models showed similar yet weaker performance, while the CNN-LSTM model produced the least accurate forecasts, with significantly higher RMSE values compared to the other models.

This study presents a robust numerical approach for solving the nonlinear Korteweg–de Vries–Burgers (KdVB) equation using the Cubic Hermite Collocation Method. The method employs piecewise cubic Hermite basis functions, which ensure both high-order accuracy and smooth derivative continuity across element boundaries. These features make the method particularly suitable for problems involving sharp gradients or smooth solution profiles. The proposed scheme is rigorously tested on a set of benchmark problems to demonstrate its effectiveness in accurately capturing the complex interplay between the dispersive and dissipative behavior inherent to the KdVB equation. Numerical results which are given with L_2 and L_∞ error norms exhibit excellent agreement with known analytical or previously published numerical solutions, and confirm the method’s stability, efficiency, and reliability. In addition, graphical representations of the numerical solutions are provided to visually illustrate the method’s performance. Due to its flexibility, accuracy, and ease of implementation, the Cubic Hermite Collocation Method proves to be a promising and efficient alternative for the numerical solution of nonlinear PDEs with mixed physical effects.

Concrete pavements in cold and de-icing environments are prone to progressive deterioration caused by freeze–thaw cycles, especially when exposed to moisture and salts. Ensuring freeze–thaw resistance is therefore critical for extending pavement service life and reducing maintenance costs. This study investigates the durability performance of concrete mixtures modified with silica fume, crumb rubber, and basalt fiber—three materials with distinct mechanisms for enhancing freeze–thaw behavior. Eight mix types—including two control groups with different water–cement ratios—were exposed to 56 freeze–thaw cycles in 3% NaCl solution and evaluated using surface scaling, mass loss, and ultrasonic pulse velocity (UPV) retention as complementary durability indicators. While the silica fume and crumb rubber blends demonstrated excellent surface resistance, the silica fume-only mix experienced complete internal degradation despite low mass loss, exposing the limitations of surface-based indicators alone. Basalt fiber-reinforced concretes showed a clear dosage-dependent improvement in internal integrity, with the 10BF and 15BF mixes retaining over 77% of their initial UPV. These results emphasize the necessity of multi-indicator durability assessments and suggest that hybrid modification strategies may offer robust protection against freeze–thaw damage in pavement-grade concretes.

In this study, complete (k,3)-arcs in the projective plane PG(2,4) that include all vertices of a complete quadrangle are systematically analyzed with respect to the inclusion of diagonal points. A computational algorithm developed in C# was employed to construct and classify such arcs. The results show that the complete quadrangle itself forms a complete (7,3)-arcs, and exactly 7560 such arcs exist depending on the choice of quadrangle point sets. Furthermore, three distinct types of complete (9,3)-arcs were identified: 24 arcs containing all four vertices and two diagonals, 48 arcs containing all four vertices and one diagonal point, and 480 arcs with all four vertices and no diagonal points. These findings reveal the combinatorial diversity of arc configurations in finite projective planes and provide new contributions to the classification of arcs. The methodology also demonstrates the effectiveness of algorithmic approaches in investigating geometric structures over finite fields.

This study presents the development of a voltammetric sensing method utilizing a glassy carbon electrode modified with electrochemically exfoliated graphene oxide (EEGO) for detecting chloramphenicol (CAP), a broad-spectrum antibiotic of concern for food safety. EEGO was synthesized through an electrochemical exfoliation process using 0.25 M LiClO4 as the electrolyte and characterized using cyclic voltammetry, scanning electron microscopy, electrochemical impedance spectroscopy, and X-ray diffraction, confirming its favorable structural and electrochemical properties. The EEGO-modified electrode exhibited superior electrochemical performance compared to bare glassy carbon electrodes, offering a broader linear range (1.0–62.5 μM) and a lower detection limit (0.067 μM) for CAP. The enhanced performance of the EEGO-modified electrode can be attributed to the high surface area, excellent electrical conductivity, and abundant oxygen-containing functional groups of EEGO, which facilitate electron transfer and promote strong analyte adsorption. The proposed sensor demonstrated excellent selectivity and stability, maintaining its performance even in the presence of common interfering substances in food matrices. The developed method was successfully applied to determine CAP in milk and honey samples, with recovery values between 84.88% and 109.48%, demonstrating its potential for practical applications in food safety monitoring. The developed voltammetric method is characterized by its simplicity and high sensitivity, eliminating the need for complex sample pretreatment. This method effectively identifies CAP in food matrices, thereby contributing to the development of practical analytical tools for monitoring food safety.