I think you probably need to check lag of other factors, a factor x1 may have low correlation, but x1(-1) or x1(-2) may significantly related to your y.

2# brvheart
谢谢你的回答 线性回归只能说明这个问题要试一试别的因子，那偏相关系数小能说明什么呢， 能直接说明这个自变量对因变量影响小吗