Effect on Paroxysmal Atrial Fibrillations of Four Predictors, 50 Patients, Traditional Versus Regularized Regressions

In this chapter we assessed the effects on paf (paroxysmal atrial fibrillations) rates of time, treatment, social and psychological scores in 50 patients. Only treatment and days of observation were significant determinants in the Traditional linear regression. The factors, social and psychological scores were unexpectly not. The best result was obtained in the Ridge regularized regression. The worst result in the standardized Traditional regression and the Lasso. The Elastic Net scored in between. The unstandardized Traditional regression produced two significant p-values (0.041 and < 0.001), despite the poor results of the standardized Traditional regression.