arXiv subjects

Tahir Cetin Akinci

Publications and source records attributed to Tahir Cetin Akinci.

6 recordsLinked to original sources

arXiv2026

Multiscale Visibility Theory: A Task-Relative Operator-Geometric Framework for Observation, Robustness, and Detectability

Multiscale Visibility Theory (MVT) is introduced as a task-relative operator-geometric framework for determining how prescribed information survives and remains accessible through a structured observation architecture. Rather than evaluating observation quality by transform magnitude, output energy, rank, or signal prominence alone, MVT measures the accessibility of a specified task direction or task subspace through the row space of the composite operator A_A = R_E W_rho H_beta P_Omega. Unlike projection-based diagnostic frameworks centered on discrimination between nominal and fault-related system behavior, MVT tracks an independently prescribed task as observation architecture, scale, representation, and retained access vary. The formulation unifies exact stagewise retention, task-subspace preservation, scale-indexed visibility and stability, worst-case erasure robustness, collective visibility, covariance-normalized statistical separation, and decision sufficiency while keeping geometric accessibility, statistical detectability, and decision adequacy mathematically distinct. A reproducible MATLAB implementation verifies the analytical relations using a controlled synthetic benchmark with a weak 260-Hz prescribed task and a stronger 380-Hz nuisance. Within the same frequency-indexed observation family, task visibility is maximized at 260 Hz, whereas complete-signal observation energy is maximized at 380 Hz. At the nominal noise level, D_max = 18 > Gamma_D = 8.563847350668, yielding a decision-visible interval of 258.5-261.5 Hz. A four-row phase-redundant observer preserves full task-model visibility through two row erasures, while two individually subthreshold branches at 258 and 262 Hz become jointly decision-visible with D_12 = 10.470833520729. All 12 numerical audits satisfy the 10^-10 tolerance.

eess.SP↗

arXiv2026

Why Model Selection Fails in Time Series Forecasting: An Empirical Study of Instability Across Data Regimes

Time series forecasting models often exhibit inconsistent performance across datasets with varying statistical and structural properties. Despite the wide range of available forecasting techniques, it remains unclear whether model selection can be reliably guided by simple data characteristics. This paper investigates why rule-based model selection fails in time series forecasting by analyzing the relationship between data-regime descriptors and model performance. A descriptor-based framework is introduced to characterize time series using measurable properties, including trend strength, seasonality, noise level, and temporal dependence. Based on these descriptors, a rule-based selection mechanism is formulated to map data regimes to candidate forecasting models. The approach is evaluated on multiple real-world datasets across different domains and forecasting horizons. The results show that rule-based model selection achieves low accuracy, with correct model identification occurring in only a small fraction of cases. Significant discrepancies are observed between recommended and empirically optimal models, particularly in noisy and mixed regimes. Further analysis reveals that model performance is highly sensitive to both dataset characteristics and forecasting horizon, resulting in substantial ranking instability across scenarios. These findings explain why simple heuristic rules fail to generalize and demonstrate that forecasting performance cannot be reliably predicted using static, descriptor-based approaches. This study provides empirical evidence that model selection in time series forecasting is inherently context-dependent and highlights the need for more adaptive, data-driven strategies.

eess.SP↗

arXiv2026

A Hybrid STFT-Based Machine Learning Framework for Physically Interpretable Arc Stability Classification in Electric Arc Welding Systems

This study presents a physically informed hybrid time-frequency and machine learning (STFT-ML) framework for arc stability monitoring in electric arc welding systems. The primary current signal is modeled as a stochastic representation of plasma dynamics and transformed into a structured feature space using localized spectral energy distributions. Within this framework, the Arc Stability Index (ASI), spectral entropy (Hs), and harmonic distortion (THDarc) are defined as energy-based descriptors and integrated with complementary time-domain features to capture both spectral redistribution and temporal variability. Experimental evaluation demonstrates that the SVM-RBF classifier achieves a hold-out accuracy of 94.4%. However, cross-validation results (85.6% for Leave-One-Out and 87.5% +/- 9.4 for 10-fold) and a 95% confidence interval of [81.65%, 92.50%] provide a more realistic assessment of generalization performance. Receiver Operating Characteristic (ROC) and Precision-Recall (PR) analyses further confirm strong class separability, particularly for stable and extinction regimes, while transient states remain more challenging due to their non-stationary nature. Compared to high-dimensional deep learning approaches, the proposed framework significantly reduces computational complexity and inference latency, enabling real-time deployment in resource-constrained environments. The results indicate that spectral energy redistribution around the fundamental frequency serves as a reliable precursor to arc instability. The main contribution of this work lies in the development of a computationally efficient and physically interpretable feature representation framework that bridges time-frequency analysis and machine learning-based classification for industrial diagnostic applications.

eess.SP↗

arXiv2026

Impact of Validation Strategy on Machine Learning Performance in EEG-Based Alcoholism Classification

Electroencephalography provides a non-invasive and cost-effective approach for analyzing neural patterns associated with alcohol dependence. However, reported classification performance in EEG-based alcoholism studies varies considerably, often due to differences in validation strategies rather than intrinsic model capability. This study presents a validation-aware machine learning framework to assess the impact of evaluation methodology on classification performance. A balanced multi-channel EEG dataset of 300 trials (150 alcoholic, 150 control) was analyzed using a structured feature representation combining statistical descriptors and spectral band interactions. Five classifiers, including support vector machines (linear and radial basis function kernels), random forest, k-nearest neighbors, and AdaBoost, were evaluated under standard and nested cross-validation protocols. Results show that conventional validation with global hyperparameter tuning introduces optimistic bias. In particular, SVM with radial basis function kernel exhibited a performance decrease of approximately 5\% under nested cross-validation, indicating overestimation. Ensemble-based methods showed more stable generalization, with AdaBoost achieving the highest performance, reaching 78.3\% accuracy ($\pm$4.25), an AUC of 0.868, and balanced sensitivity (78.67\%) and specificity (81.33\%). These findings highlight that validation strategy is a primary determinant of perceived model performance. Statistical analysis using McNemar's test further shows that most performance differences between models are not statistically significant, emphasizing careful interpretation of classification results. The proposed framework provides a reproducible and robust basis for evaluating machine learning models in biomedical signal analysis.

eess.SP↗

arXiv2026

Operator-Theoretic Energy Functionals for Impulse-Excited Nonstationary Signal Analysis

This study presents an operator theoretic framework for defect detection in impulse excited nonstationary systems. Measured responses are modeled as finite energy impulse responses perturbed by stochastic disturbances and represented in the Hilbert space L2(R). Time frequency representations are formulated as bounded linear analysis operators associated with continuous frames, enabling a consistent description of how structural perturbations redistribute transient signal energy. Within this formulation, a nonlinear Energy Concentration Index ECI is introduced to quantify localized transform domain energy over selected regions of the time frequency plane. The boundedness and continuity of the functional ensure that small physical variations in system parameters produce measurable changes in localized energy distribution. This property enables the construction of a statistical separability functional that links multi resolution energy geometry to classification performance. Based on these results, a compact Impulse Based Multi Resolution Energy Detector IMRED is derived. The analysis shows that variations in damping and resonant frequency produce systematic changes in time frequency coefficients and localized energy concentration. Experimental validation using impulse excited ceramic measurements demonstrates that the proposed descriptor captures defect induced structural differences with strong discriminative capability. The resulting IMRED statistic achieves an AUC of 0.908 and provides clearer class separation than global Fourier band energy measures and non optimized wavelet band aggregation. These results establish a direct relationship between impulse response modeling, localized energy geometry, and statistical decision mechanisms, providing a mathematically grounded basis for energy driven defect detection in structural monitoring applications.

eess.SP↗

arXiv2019

ROSETLineBot: One-Wheel-Drive Low-Cost Power Line Inspection Robot Design and Control

Continuous operation of the electrical transmission lines is of great importance for today's electricity-dependent world. It is crucial to detect and diagnose a possible problem on the lines before that occurs. In general, maintenance, repair and error detection operations of the electrical transmission lines are performed by humans periodically. However, these difficult tasks contain high-risks in terms of occupational health and safety. Moreover, it is not possible to continuously inspect these lines which extend over long distances by maintenance and repair personnel. For all these reasons, it is very convenient to use robotic systems for inspection on power transmission lines. By evaluating the data which are provided from sensors and cameras of robotic systems, it would be possible to take precautions before an error occurs. Robotic systems that can move continuously along the line would be able to gather continuous and uninterrupted information. Thus, potential problems can be detected and identified in advance. In this study, a single wheel drive low-cost mobile robot is designed and controlled which can work and move on the electrical transmission lines. Because of the modular structure of the designed robot, different types of sensors can be integrated into the robot, easily. The designed low-cost robotic system will cause minimum losses in case of possible breakage.

eess.SY↗