© 2023, The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.Electroencephalography (EEG) signals are named letters of the brain, and their translation is a complex issue. This work recommends a new hand-crafted feature-based EEG signal classification model, including a new local histogram-based feature generation function, the cube pattern. The recommended model comprises preprocessing/signal denoising, feature extraction using the presented cube pattern, neighborhood component analysis-based feature selection, and classification by employing 25 classifiers. Multi-scale principal component analysis (MSPCA) is applied to the raw EEG signals in the denoising phase. Afterward, the denoised EEG signals are forwarded to the feature extraction method. Next, tunable q-factor wavelet transform (TQWT) is employed to denoise signals for decomposition, and levels/sub-bands are generated. The selected features are classified from 25 classifiers by using the MATLAB Classification Learning tool. The presented model is applied to a commonly used EEG signal dataset. Variable performance evaluation metrics are used to test the performance of each classifier. Per the calculated results, the presented model reached over 99% accuracy using 24 of the 25 classifiers, and a comprehensive benchmark is obtained. The calculated results and obtained findings denote the high performance of the presented cube pattern and the neighborhood component analysis-based model.