Renmin Zhujiang (Jan 2024)
Carrying Capacity Evaluation on Water Resources of Jilin Province Based on PCA-GA-XGboost Model
Abstract
To improve the efficiency and accuracy of carrying capacity evaluation in water resources, this paper proposes an indicator evaluation model based on principal component analysis (PCA), genetic algorithm (GA), and eXtreme gradient boosting tree (XGBoost).Meanwhile, fourteen evaluation indicators are defined with water resources, socio-economics, and ecological environment employed as subsystems.PCA is adopted to reduce the dimensionality of the evaluation indicators.Additionally, based on XGBoost, this paper conducts an evaluation analysis on the carrying capacity of water resources from 2011 to 2021 and utilizes GA to optimize four parameters in XGBoost.The results show that after simplifying the evaluation indicators by PCA, the correlation coefficient of the PCA-GA-XGBoost model is better than GA-B, GA-SVM, GA-XGBoost, and XGBoost.The carrying capacity of water resources in Jilin Province from 2011 to 2021 is between 0.192 and 0.724, presenting a trend of first increasing, then decreasing, and finally increasing with improved carrying capacity situation.Meanwhile, the built-in function of eigenvalue importance ranking in the model is leveraged to conclude the fact that the indicator with the largest importance is identified as the applied fertilizer amount per hectare (0.530 7).
