%0 Journal Article
%T 基于ANFIS模型的草原土壤湿度预测<br>Grassland Soil Moisture Prediction Based on ANFIS Model
%A 李帅
%A 陈成
%A 马明明
%A 王自强
%J Modeling and Simulation
%P 1481-1490
%@ 2324-870X
%D 2023
%I Hans Publishing
%R 10.12677/MOS.2023.122138
%X 中国是一个资源大国，草地资源十分丰富，面积分布也在世界前列。草原生态系统不仅是维护我国生态系统稳定的重要屏障，同时还为我国的经济发展提供保障。近年来，快速发展的畜牧业使得草地退化严重，有的甚至出现了沙化现象。提供科学的草地管理方式迫在眉睫，故对土壤中的湿度预测对于草原的保护和开发具有重要的意义。本文对往年的统计数据进行分析，然后通过数学模型对2022、2023年的土壤湿度进行预测。首先对数据进行共线性分析，对于共线性强的数据采用Lasso回归的方法进行降维。之后用ARIMA时间序列方法对以往年月数据进行预测。最后建立输入(Lasso回归所筛选变量和土壤蒸发量变量)和输出(不同深度土壤湿度) ANFIS模型，对往年所测数据进行训练且整体数据集合训练拟合度均在85%以上，该模型的准确度较高。最后通过训练好的ANFIS模型预测2022年、2023年不同深度土壤湿度。<br />
China is a big resource country. Grassland resources are very rich and the area distribution is also among the top in the world. Grassland ecosystem is not only an important barrier for maintaining the stability of Chinese ecosystem, but also provides a guarantee for the economic development of our country. In recent years, the rapid development of animal husbandry has caused serious deg-radation of grassland, and some even appear desertification phenomenon. It is urgent to provide scientific grassland management methods, so soil moisture prediction is of great significance for grassland protection and development. This paper analyzes the statistical data of previous years, and then predicts the soil moisture in 2022, 2023 through mathematical models. Firstly, collinear-ity analysis was carried out on the data, and dimension reduction was carried out on the data with strong collinearity by Lasso regression method. Then the ARIMA time series method is used to pre-dict the data of previous months. Finally, ANFIS models for input (variables screened by Lasso re-gression and soil evaporation variables) and output (soil moisture at different depths) were estab-lished. The data measured in previous years were trained and the overall data set training fit de-gree was above 85%, indicating a high accuracy of this model. Finally, the trained ANFIS model was used to predict the soil moisture at different depths in 2022 and 2023.
%K Lasso回归，ARIMA时间序列方法，ANFIS模型，土壤湿度；Lasso Return
%K ARIMA Time Series Method
%K ANFIS Model
%K Soil Moisture
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=63080