Platform for project evaluation based on soft-computing techniques

Abstract

Project management tools require a high added value. However, most standards do not refer to the use of techniques that treat uncertain information. This paper presents experiences in project evaluation process by using soft computing techniques. We present a project management tool based on LABPRO Software Ecosystem, which combines R functionalities with postgres database system. LABPRO is a Software Ecosystem which contains tools for making decisions in project management organizations, and implements processes of PMBOOK standard. We present the results of applied different soft computing algorithms, published in FRBS package, and included in LABPRO platform. We compare algorithms Anfis, Hyfis, FS.HGD based on neural networks with the algorithms GFS LS RS, GFS THRIFT based on genetic algorithm techniques. The algorithms selected learn from a database that contains projects already evaluated, and after that, they classify other projects. We applied cross validation techniques combined with Friedman test and Wilcoxon test. Finally, we present the comparison results and the conclusions.