We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
不同训练时间跨度情况下 3 种机器 学习模型估算 ET<sub>0</sub> 性能研究.
- Authors
李思颖
- Abstract
In order to explore the performance of three machine learning models in estimating ET0 under different training time spans, this paper collected the meteorological data of five meteorological stations in the arid area of China, namely Erenhot, Dulan, Golmud, Hami and Turpan, from 1960 to 2019, and aimed at the seven climatic factors of maximum temperature, minimum temperature, average temperature, relative humidity, wind speed, daily extraterrestrial solar radiation and global solar radiation, Taking the results of daily ET0 calculated by PM formula as a standard, this paper discusses the prediction of daily ET0 by AdaBoost, XGBoost and RF machine learning models under different training set time spans. The results show that: ① the comparison of estimation accuracy of the three machine learning models under different training time spans: AdaBoost model > RF model > XGBoost model. ②From the perspective of the data volume of the training set, it is recommended to take 20 years as the duration of the training set to estimate the daily ET0, so as to achieve the prediction effect that is longer than the duration of the training set.
- Subjects
ASTROPHYSICAL radiation; GLOBAL radiation; SOLAR radiation; METEOROLOGICAL stations; MACHINE learning; WIND speed
- Publication
Water Conservancy Science & Techonlogy & Economy, 2023, Vol 29, Issue 4, p104
- ISSN
1006-7175
- Publication type
Article
- DOI
10.3969/j.issn.1006-7175.2023.04.022