TY - JOUR AU - Shengping Zhang, AU - Qi, Jie PY - 2023/05/02 Y2 - 2024/03/29 TI - A STUDY ON THE RELIABILITY OF AN ARTIFICIAL INTELLIGENCE MODEL FOR WATER ENVIRONMENT EVALUATION JF - GEOMATE Journal JA - INTERNATIONAL JOURNAL OF GEOMATE VL - 25 IS - 107 SE - Articles DO - UR - https://geomatejournal.com/geomate/article/view/3915 SP - 220-227 AB - <p>The objective of this study is to discover the most effective water environment improvement measures for the 109 most important watersheds of Japan, which are well-known Class-A watersheds under the jurisdiction of the Japanese government. An artificial intelligence (AI) model has been created by applying Deep Learning technologies with the expectation that an AI model is able to take adevantage of the multiple categorical water environmental data, and watershed information from 109 watersheds has been collected as teacher data to train the AI model. This study aims to find the best way to present the water environment data to an artificial intelligence model. To provide the most reliable water quality estimations, three different ways of water environment data presentation have been examined. It has been identified that presenting the raw environment data to the AI model as teacher data is the best way for building an AI model. This study concludes by pointing out that preprocessing data will cause information loss of teacher data and will also add biased information to teacher data. Therefore, using raw data without preprocessing as teacher data will help build a more reliable AI model. It is hoped that this study will contribute to establishing a more reliable river environment planning and management methodology.</p> ER -