Methods for regression analysis in high-dimensional data
چکیده(به انگلیسی؛ چکیدهی فارسی در دسترس نیست)
By evolving science, knowledge and technology, new and precise methods for measuring, collecting and recording information have been innovated, which have resulted in the appearance and development of high-dimensional data. The high-dimensional data set, i.e., a data set in which the number of explanatory variables is much larger than the number of observations, cannot be easily analyzed by traditional and classical methods, same as the ordinary least-squares method, and its interpretability will be very complex. Although, in classical regression analysis, the ordinary least-squares estimation is the best estimation method if the essential assumptions are met, but it is not applicable for high-dimensional data and in this cconditions, we need to apply the modern methods. In this research, it is firstly mentioned to the drawbacks of classical methods in analysis of high-dimensional data and then, it is proceeded to introduce and explain about the modern and common approaches of the regression analysis for high-dimensional data same as principal component analysis and penalized methods. Finally, a simulation study is performed to apply and compare the mentioned methods in high-dimensional data.
نویسندگان و وابستگی سازمانی
- Maanavi, Monireh — Semnan University
- Roozbeh, Mahdi — Semnan University