Adj.R-Square是什么
发布网友
共3个回答
热心网友
R-Square是变量之间相关程度的值。
在多元回归分析中，Adj. R-Square根据样本数量和自由度调整后的样本R-Square。考虑了自变量(变量)数目的影响。
热心网友
R squared=1-SSE/SST,英文叫Coefficient of determination
The numerator of the fraction of r formula is
sum[(x-mean(x))*(y-mean(y))],and to counterbalance the dispersions of x and y, the denominator is std(x) and std(y). So r ranges from -1 to 1, and the greater is the absolute value of r, the closer of the ordered-pairs points to a line (the regression line). With a positive/negative r, the regression line has a positive/negative slope.
And to put the focus on the linearity between the two ramdom variables, people use the square of r more than r itself and it now ranges from 0 to 1. The extreme of 1 occurs when all the points are exactly on a sloppy line (which almost wouldn't happen to the real statisticcal application), and the other extreme 0 reflects that they don't show a line pattern at all, but it doesn't exclude the possibility that that are still not independent. Say, when the distribution pattern are symmetric in a vertical or horizental line.
热心网友
Adj. R-Square等信息，没有相关系数R或者R^2以及标准偏差SD、检验值F等在你作图用的数据所在的workbook里面，下面sheet2里面有拟合的详细参数
