New linguistic data summarization approach in predictions problems in a project management application

Abstract

In the proposal, multilayer networks are built where the layers represent the different elements that make up the linguistic summaries and systems with single–value triangular neutrosophic numbers are combined for the computation of the inference. In the validation, the new inference model was applied in the scenario of making–decision in project management and the “Arithmetic mean of the error in the prediction” was applied as a metric. Comparisons are made by evaluating the inference model with different abstract parameters. In addition, the linguistic summaries are combined with the fuzzy cognitive maps and the extension NCM_LDS is generated. The new proposed extension is compared with other map extensions, it is shown that the proposed inference mode–lo reports better results than the rest of the maps in the diagnostic stage, but that it does not have significantly better results than the m–FCM in the decision and prognosis stages. Pág. 93The NCM_Indeterminacy map is shown to be the one with the worst results.