Explainable Artificial Intelligence with Single Layer Feedforward Neural Network and Improved Crowned Porcupine Optimization Algorithm for Classification Problems DOI Open Access

S Caxton Emerald,

T. Vengattaraman

Engineering Technology & Applied Science Research, Год журнала: 2025, Номер 15(2), С. 21593 - 21598

Опубликована: Апрель 3, 2025

The increasing occurrence of network intrusions calls for the development advanced Artificial Intelligence (AI) techniques to tackle classification challenges in Intrusion Detection Systems (IDSs). However, complex decision-making processes AI often prevent human security professionals from fully understanding behavior model. Explainable (XAI) enhances trust IDSs by providing transparency and assisting interpreting data reasoning. This study explores that improve both accuracy interpretability, strengthening management cybersecurity. Integrating performance with explainability improves builds confidence automated systems classifying intrusions. presents an Kernel Extreme Learning Machine Improved Crowned Porcupine Optimization Algorithm (XAIKELM-ICPOA) approach. Initially, proposed XAIKELM-ICPOA method preprocesses using min-max scaling ensure uniformity model performance. Next, (KELM) is employed classification. (ICPO) used optimize KELM hyperparameters, improving Finally, SHAP as XAI technique provide insights into feature contributions processes. was evaluated on NSL-KDD dataset, achieving 96.82%.

Язык: Английский

Explainable Artificial Intelligence with Single Layer Feedforward Neural Network and Improved Crowned Porcupine Optimization Algorithm for Classification Problems DOI Open Access

S Caxton Emerald,

T. Vengattaraman

Engineering Technology & Applied Science Research, Год журнала: 2025, Номер 15(2), С. 21593 - 21598

Опубликована: Апрель 3, 2025

The increasing occurrence of network intrusions calls for the development advanced Artificial Intelligence (AI) techniques to tackle classification challenges in Intrusion Detection Systems (IDSs). However, complex decision-making processes AI often prevent human security professionals from fully understanding behavior model. Explainable (XAI) enhances trust IDSs by providing transparency and assisting interpreting data reasoning. This study explores that improve both accuracy interpretability, strengthening management cybersecurity. Integrating performance with explainability improves builds confidence automated systems classifying intrusions. presents an Kernel Extreme Learning Machine Improved Crowned Porcupine Optimization Algorithm (XAIKELM-ICPOA) approach. Initially, proposed XAIKELM-ICPOA method preprocesses using min-max scaling ensure uniformity model performance. Next, (KELM) is employed classification. (ICPO) used optimize KELM hyperparameters, improving Finally, SHAP as XAI technique provide insights into feature contributions processes. was evaluated on NSL-KDD dataset, achieving 96.82%.

Язык: Английский

Процитировано

0