A new mathematical modeling method based on space geometry theory

Regression analysis is a kind of mathematical modeling approach to find the relation which hidden in variables through experiment and observation. This method can describe a perfect relation in many problems usually. But the amount of computation is high because we must use a complexity mathematical expression when wanted to get a satisfying result. This paper analyzes mathematical behavior of linear regression in high dimensional space, and finds the relation between variables and their functions with dependent variables in measurement space. Experiment proved the approach is well-defined, low computation complexity and effective.