Similarity or distance measures are core components used by distance-based clustering algorithms to cluster similar data points into the same clusters, while dissimilar or distant data points are placed into different clusters. The performance of similarity measures is mostly addressed in two or three-dimensional spaces, beyond which, to the best of our knowledge, there is no empirical study that has revealed the behavior of similarity measures when dealing with high-dimensional datasets. To fill this gap, a technical framework is proposed in this study to analyze, compare and benchmark the influence of different similarity measures on the results of distance-based clustering algorithms. For reproducibility purposes, fifteen publicly available datasets were used for this study, and consequently, future distance measures can be evaluated and compared with the results of the measures discussed in this work. These datasets were classified as low and high-dimensional categories to study the performance of each measure against each category. This research should help the research community to identify suitable distance measures for datasets and also to facilitate a comparison and evaluation of the newly proposed similarity or distance measures with traditional ones.

Clustering neurodata, including electroencephalography (EEG) signals, is crucial for brain-computer interface (BCI) and neurological analysis. However, traditional methods struggle with noise, overlapping distributions, and high-dimensional data. This study presents the Credibility-Adjusted Data Conscious Clustering Method (CADCCM), an adaptive computational learning model for clustering neurodata, including EEG signals. CADCCM improves clustering robustness by dynamically adjusting assignments through credibility updates and optimizing parameters for better accuracy and stability in noisy, high-dimensional data. CADCCM dynamically adjusts cluster assignments by integrating alpha and beta parameters to balance fuzzy membership and credibility. A grid search framework optimizes clustering parameters, and preprocessing techniques (Fourier Transform, Wavelet Transform, and Gaussian filtering) improve feature separability. The method is benchmarked against traditional and recent Clustering methods across 13 datasets. CADCCM achieves superior clustering performance, consistently outperforming baseline methods in Rand Index (RI), F-score, and Cohen’s Kappa, particularly in noisy datasets. Gaussian filtering further enhances clustering accuracy. GPU acceleration ensures computational feasibility for large-scale neurodata applications. Additionally, CADCCM outperforms other methods by satisfying all clustering properties, including scale invariance, richness, consistency, and order independence. CADCCM bridges the gap between traditional and advanced clustering by introducing credibility-based updates and optimized parameter selection, leading to improved accuracy, robustness, and efficiency. The method holds promise for applications in cognitive state monitoring, neurological disorder detection, and BCI systems. CADCCM enhances the interpretability, reliability, and scalability of clustering. Future research will focus on real-time clustering, adaptive hyperparameter tuning, and deep-learning-based feature extraction.