heuristic methods to combat the regression challenges

Nowadays, high–dimensional data in which the number of observations issmaller than the number of parameters, appear in many practical applications such asbiosciences, social networks, psychological researches, recommendation systems and soon. in the regression model analysis, the well–known ordinary least–squares estimationmay not be applicable when the classical assumptions such as normality of theerror terms and full ranking of the design matrix are violated. as known, a successfulapproach for high–dimension cases is the penalized scheme (such as lasso) with theaim of obtaining a subset of effective explanatory variables that predict the dependentvariable as the best, while setting the other parameters to zero. here, we review anddevelop several penalized models to be used in high-dimension regression analysis forhigh-dimensional data sets. in this paper, we apply eye data to evaluating a strategyfor detecting human eye disease.