Linear regression algorithm will encounter two problems in selected variables: First, remove multicollinearity interference, the second is to select the optimal combination of independent variables.
Linear regression step
1. Selection of variables
Note removed point multicollinearity interference, selecting an optimal combination of independent variables. It should be appreciated that the coefficient of determination: R ^. It is the understanding of the variables selected from two basic problems.
2. Create a line regression model
3. Analysis Model
R^
It represents the dependent variable being the fluctuation model fitting percentage, calculated to measure data quality of model fit.
Mathematical formula defined
Common R ^ recommended in a single arguments
Adjustment R ^
When the input argument has more than one model, we have to make adjustments to the R ^, this time it is called adjusted R ^
Adjusted R ^ recommended for use in multi-arguments.
Based on a linear regression model to understand the meaning R ^ ranges indicated
Definition appreciated the need to adjust a variety of independent variables based on the model of the R ^
Join an unrelated argument in the original data, R ^ will subsequently be misled increases. The figure is increased by one, and the coin toss independent data results in the original data.
The value range of experience to judge