Enhancing Car Segmentation for Thailand's Expressway Industry With an Automated Hybrid Machine Learning Framework

Abstract

Car segmentation on Thailand's expressways poses challenges for traditional models due to unique characteristics, often resulting in predictive inaccuracies. Manual data analytics in this field is time-consuming and human centric. This research introduces an Automated Hybrid Machine Learning (AHML) framework leveraging advancements in AutoML, tailored for personalized customer segmentation in Thailand's expressway industry. The framework streamlines and automates the machine learning process, aiming to expedite model construction while enhancing performance. By employing clustering as an initial step followed by the Random Forest classifier as a hybrid classification approach, significant performance improvements are achieved compared to existing methods. Specifically, the model outperforms by 9.15% and 12.84% in both clusters, respectively. This research highlights the potential of the framework to address complex segmentation challenges and advance personalized customer targeting.