Feature Engineering: The Impact on Artificial Intelligence Model Performance

Feature engineering is a crucial component of AI model development, significantly impacting performance, interpretability, and scalability. There are two types of feature engineering, the traditional and automated feature engineering, each has its own strengths and challenges. In order to improve model accuracy and interpretability, traditional feature engineering manually selects and creates features using domain expertise. However, it is time-consuming and might not work well with large, complex datasets. In contrast, automated feature engineering is scalable and efficient, sometimes at the expense of interpretability, since it uses sophisticated algorithms and tools to rapidly generate and refine features. Various techniques, including domain knowledge-based features and statistical features, are discussed. The advantages of feature engineering are demonstrated by case studies that include from the fields of security, healthcare, and natural language processing. Although there are still issues with scalability, the need for domain expertise, and overfitting, feature engineering is progressing and paving the way to more effective methods that drive innovations in AI model accuracy and applicability.

Biography

MOHD HANAFI AHMAD HIJAZI (M’14) received the B.Sc. and M.Sc. degrees in computer science at the Universiti Teknologi Malaysia, in 2001 and 2005 and the Ph.D. degree in computer science from the University of Liverpool, United Kingdom, in 2012. From 2012 to 2018 he was a Senior Lecturer with the Faculty of Computing and Informatics, Universiti Malaysia Sabah. Since 2018, he has been an Associate Professor at the same Faculty. Currently, he is the Dean of the faculty and the Head of Data Technology and Applications Research Group. His research interest includes data mining and artificial intelligence with the application on healthcare and biometrics. Dr. Hijazi is also a professional member of Malaysia Board of Technologist (MBOT) and has several times been appointed as auditor/ lead auditor for the accreditation of IT programmes under the MBOT.