AN ARTIFICIAL INTELLIGENCE MODEL FOR IDENTIFYING RIVER ENVIRONMENT IMPROVEMENT MEASURES

Abstract

The objective of this study is to identify the most effective measures to improve the environments of Japan's 109 primary watersheds, designated as Class-A Watersheds, which are overseen by the national government. Traditionally, the selection of environmental improvement measures for each watershed has relied heavily on simulations from analytical models and the personal insights of river management officers or environmental planners familiar with specific watersheds. However, these analytical models are constrained by their limited capacity to incorporate a finite array of quantifiable factors, and river management officers or environmental planners, despite their profound expertise in particular rivers, may not possess extensive experience with all significant rivers within a region. This study first designed an artificial intelligence model to evaluate river environments, using a comprehensive dataset that includes both numerical and categorical data for the 109 Class-A watersheds. Furthermore, the reliability of the artificial intelligence model has been verified and applied to identify the most effective environmental improvement measures for each watershed. The study concluded by indicating that the artificial intelligence model is both reliable and useful in determining which improvements will effectively enhance the river environments. This study is anticipated to contribute to the establishment of a more robust and reliable methodology for river environment planning and management.