That is why data transformation prior to applying our model is very important in this case. Many variables do not meet assumptions of linear relationship, normality, independence and homoscedasticity of error terms when used on linear model or in statistical tests, these may give misleading results. ![]() Perform transformation on the data so that the linear regression model works for that transformed data. To let go of the linear regression model and adopt a more suitable model or 2. In our case if a data does not fit over a linear regression model we have two basic choices 1. The purpose of transforming data is to make the data follow assumptions of statistical inference or undergo a parametrical statistical test or fit over a model. In statistics, data transformation is the application of a deterministic mathematical function each point in a data set - that is, each data point zi is replaced with the transformed value y = f(zi), where f is a function. ![]() It also throws light on using alternate generalized linear models that would allow for flexibility in terms of getting normalized error terms. This article discusses transformations in statistics and does a comparative study of different data transformation techniques used while working on linear regression models.
0 Comments
Leave a Reply. |