请用英文翻译：在任务2中，我们首先分析了四类土地的特点。随后我们对数据进行了标准化处理与主成分分析。之后，我们进行了KMO检验、Barttle球状检验的共线性检验，来判断主成分分析的可行性。在主成分分析的基础上，我们选用BP神经网络算法，加权求出新的变量。经过多次训练后，发现回归效果显著，以此为基础，针对四个典型地区做出了相应权重的修正。

In Task 2, we first analyzed the characteristics of the four types of land. Then we standardized the data and performed principal component analysis. Afterwards, we conducted KMO test and Barttle sphere test to assess the feasibility of principal component analysis. Based on the principal component analysis, we selected BP neural network algorithm to calculate the new variables with weighting. After multiple trainings, we found that the regression effect was significant. On this basis, we made corresponding weight corrections for the four typical regions.