Table Final total influence matrix Factor
At this stage, a heuristic algorithm has been proposed to provide a full ranking of organization. First, we should select the best FDEA model at a given level of α-cut. In this regard, the preferred FDEA model (or best level of α-cut) is determined based on correlation coefficients calculated for FDEA models at five different levels of α-cut and DEA and FCM models. Selected model is a model that has the maximum correlation coefficient with other two models. Thus, based on tests carried out, it LBH589 is shown that FCM, DEA and FDEA methods have the highest correlation coefficients with each other at α-cut = 0.7. The performance scores of these three models are normalized (i.e., all performance values are converted values between 0 and 1). Then, the average normalized performance scores are obtained from DEA and FCM and FDEA (at α-cut = 0.7) for each organization are calculated. Therefore, a full ranking of organizations is provided based on the average values. The results of lateral roots algorithm are presented in Table 9.