通过已知数据建立对应模型，找出数据集的最优线性关系来预测或估计需要的因变量。(The corresponding model is established by the known data, and the optimal linear relation of the data set is found to predict or estimate the dependent variable.)