urban carbon emissions 1992-2021 in China

This dataset was obtained by fitting a gradient boosted regression tree (GBRT).Before using the GBRT model, k-means clustering was used to classify China into two major categories, and then regression was fitted to the carbon emissions of the two categories of cities separately, where the overall fitted R2 for the first category of cities was 0.866, and the overall fitted accuracy for the second category of cities was 0.974.It is important to note that these values represent carbon emissions per square kilometre per city not CO2 emissions, and if you need to calculate CO2 emissions you need (*44/12*area).

本数据集通过梯度提升回归树（Gradient Boosted Regression Tree, GBRT）拟合生成。在使用该GBRT模型前，先采用k-means聚类算法将中国的城市划分为两大类别，随后分别针对两类城市的碳排放数据开展独立回归拟合。其中第一类城市的整体拟合R²为0.866，第二类城市的整体拟合精度为0.974。需特别说明的是，上述数值代表的是各城市每平方公里的碳排放量，而非二氧化碳（CO₂）排放量；若需计算二氧化碳排放量，需按公式（*44/12*面积）进行转换。