Unveiling the impact of machine learning algorithms on the quality of online geocoding services: a case study using COVID-19 data

• DOI: 10.1007/s10109-023-00435-8
• Publication Date: 2024-01-25T07:02:23Z
• Authors: Batuhan Kilic; Onur Can Bayrak; Fatih Gülgen; Mert Gurturk; Perihan Abay
• Keywords: Geocoding

Abstract In today's era, the address plays a crucial role as one of key components that enable mobility in daily life. Address data are used by global map platforms and location-based services to pinpoint geographically referenced location. Geocoding provided online is useful spatial tracking reported cases controls analysis infectious illnesses such COVID-19. The first most critical phase geocoding process matching. However, due typographical errors, variations abbreviations used, incomplete or malformed addresses, matching can seldom be performed with 100% accuracy. purpose this research examine capabilities machine learning classifiers measure consistency results produced identify best performing classifier. performance seven was compared using several text similarity measures, which assess match scores between input services' output. utilized testing came from four distinct applied 925 addresses Türkiye. findings study revealed Random Forest classifier accurate procedure. While hold true for similar datasets Türkiye, additional required determine whether they apply other countries.