COMPARATIVE STUDY OF DIFFERENT DECISION TREE ALGORITHMS FOR PV SUITABILITY EVALUATION IN QINGHAI PLATEAU
Taking the Qinghai Plateau as an example,a total of 185 photovoltaic sites positional information are collected through field investigation and integration of Google Images.Based on this dataset,three machine learning algorithms,namely Classification and Regression Tree(CART),Random Forest(RF),and Extreme Gradient Boosting(XGBoost),are compared and evaluated for their predictive performance is assesed using ROC curves and statistical indicators.The results reveal that XGBoost demonstrates superior predictive performance and robust adaptability to noisy data,overall outperforms the other models.Factors such as total solar radiation,distance from the power grid,and distance to roads are identified as the key factors influencing the location of photovoltaic power stations.The PV suitability maps generated by the three models indicate that the highly suitable areas are primarily distributed in the Qaidam Basin and Gonghe Basin.The highly suitable and relatively suitable areas account for 15.31% and 16.33% of the total area of the study area,respectively.
photovoltaic power stationzoningresource valuationmachine learningQinghai PlateauArcGIS
NETL
NSTL
万方数据
