It is the regression analysis is the foundation stone of knowledge of statistics , which mostly depends on the ordinary least square method , but as is well known that the way the above mentioned her several conditions to operate accurately and the results can be unreliable , add to that the lack of certain conditions make it impossible to complete the work and analysis method and among those conditions are the multi-co linearity problem , and we are in the process of detected that problem between the independent variables using farrar –glauber test , in addition to the requirement linearity data and the lack of the condition last has been resorting to the nonparametric regression and processor the problem using kernel ridge regression function and that depend on estimate band width ( smoothing parameter ) therefore has been resorting to two different ways to estimate the parameter and are Rule of thumb (RULE) and Bootstrap (BOOT) and comparison between those ways using the style of simulation
