System Modeling of AkT using Linear and Robust Regression Analysis

This paper presents the review of different types of modeling like deterministic and probabilistic. Main stress of this paper is on Probabilistic modeling. Probabilistic modeling is further divided into three types: predictive, regression and correlation modeling. The analysis was done using three different input proteins and four different output proteins. A model was made using different analysis techniques which presents the r² values for one of the marker protein called AkT. This paper investigates two linear regression models: ordinary least square analysis and partial least square analysis. In this paper we are also comparing different robust regression analysis techniques with partial least square method and plotted the residual and predicted values of the different techniques, PLS yields the best result in terms of squared R and plots. For all these calculations we have used Systat 13 software.