Voice pathology detection using machine learning algorithms based on different voice databases

•Dual-Database Evaluation: Unlike prior research that predominantly trains and tests algorithms on a single database, this work explores the robustness of machine learning algorithms through cross-database evaluations. This is critical for assessing the generalizability of algorithms when applied to real-world scenarios.•Balanced Databases: The use of balanced datasets ensures equitable representation of both healthy and pathological voice samples, minimizing biases during training and evaluation.•Voice sample Selection: by focusing on the vowel /a/ in both databases, the study ensures consistency in feature extraction and minimizes variability due to phonetic differences. Moreover, the vowel /a/ is particularly significant in voice pathology analysis as it originates directly from the larynx without using mouth organs.•Evaluation Metrics: Beyond accuracy, the study employs precision, sensitivity, specificity, F-measure, and G-mean to provide a holistic evaluation of algorithm performance.•Cross-Database Testing: Scenario 2, where algorithms are trained on one database and tested on another, significantly contributes to understanding the reliability of machine learning systems in handling new and diverse datasets. In other words, it is highly imperative to evaluate the generalizability and robustness of machine learning models in scenarios where data originates from different environments or sources.•Superior performance of OSELM: The consistently high performance of the OSELM algorithm, especially in cross-database evaluations, demonstrates its potential as a robust tool for voice pathology detection in practical applications. The application of machine learning in analyzing voice disorders has become crucial for non-invasive voice pathology detection using voice signals. However, current systems face challenges such as low detection accuracy, limited databases, and evaluation metrics. More importantly, most existing studies rely on training and testing algorithms based on the same database, limiting their applicability in real-world scenarios with diverse data sources. Unlike traditional approaches that focus solely on single-database training and testing, this study presents a cross-database evaluation strategy to assess the robustness and generalizability of machine learning algorithms for voice pathology detection. Several algorithms, including Online Sequential Extreme Learning Machine (OSELM), Support Vector Machine (SVM), Decision Tree (DT), and Na