Logistic Regression Model of Built-Up Land Based on Grid-Digitized Data Structure: A Case Study of Krabi, Thailand

In order to measure and detect land-use changes, it is necessary to know the pace at which land-use changes from one type of land to another. However, due to limited resources, researchers are having difficulty doing statistical analyses on polygonal data structures. The digital data structure lends itself to statistical analysis using general-purpose software, whereas land-use change was assessed solely by counting grid cells. The polygonal data were converted to digital data using the grid-digitized approach. This study compares land-use changes in Thailand's Krabi province in 2000–2009, and 2009–2018. Thematic maps and the bubble plot were used to depict land-use change across Krabi province, with a digitized grid (100 by 100 m) encompassing the entire region. A logistic regression model was used to examine the probability of built-up land. According to the findings, total built-up land was 5164 hectares (1.1% of all areas) in 2000, 9881 hectares (2.1% of all areas) in 2009, and 18,832 hectares in 2018. (4% of all areas). Built-up land rose by 91.3% and 90.6%, respectively, between 2009 and 2018. Each location's land-use was related to the probability of built-up land. Rubber plantation and agricultural land provided the majority of the built-up land. Furthermore, areas closer to metropolitan centers saw a higher percentage rise in built-up land than rural areas. Receiver operating characteristic curves and F-scores indicated that the models were accurate enough.