基于机器学习的土壤中营养元素预测:以N元素为例  

Prediction of nutrient elements in soil based on machine learning:taking Nitrogen as examples

作　　者：方景文 彭玉旭[2] 郭军 卢欣 FANG Jingwen;PENG Yuxu;GUO Jun;LU Xin(Changsha Natural Resources Comprehensive Survey Center of China Geological Survey,Changsha 410600,Hunan,China;School of Computer and Communication Engineering,Changsha University of Science&Technology,Changsha 410600,Hunan,China)

机构地区：[1]中国地质调查局长沙自然资源综合调查中心,湖南长沙410600 [2]长沙理工大学计算机与通信工程学院,湖南长沙410600

出　　处：《资源信息与工程》2024年第4期73-75,共3页Resource Information and Engineering

基　　金：中国地质调查局项目(DD20208071);中国地质调查局项目(DD20230621)。

摘　　要：本文研究了基于BP神经网络算法的土地质量调查中地球化学元素含量的预测方法,通过岳阳市耕地区表层土壤地球化学元素含量的调查,利用已知的元素含量数据建立机器学习模型,实现对未知元素含量的准确预测。本文采用SelectKBest函数进行变量选择,并通过训练集、验证集和外部测试集的数据评估模型性能。结果表明,所建立的模型具有较高的预测精度和泛化能力,为土地质量调查中地球化学元素含量的预测提供了一种便捷有效的方法。This article studied the prediction method of geochemical element content in land quality survey based on BP neural network algorithm.By investigating the geochemical element content of surface soil in the cultivated area of Yueyang City,a machine learning model was established using known element content data to accurately predict the unknown element content.This article utilized the SelectKBest function for variable selection and evaluated the model performance through data from the training set,validation set,and external test set.The results indicate that the established model has high prediction accuracy and generalization ability,providing a convenient and effective method for predicting geochemical element content in land quality surveys.

关 键 词：机器学习 BP神经网络 表层土壤 元素预测 

分 类 号：P642[天文地球—工程地质学]