Development of a machine learning-based tool for predicting participants in evaluation processes
selection processes; logistical efficiency; machine learning; data forecasting; decision support systems.
This work was developed within the Núcleo Permanente de Concursos (Permanent Center of Contests) of UFRN, the Comperve. All the management of the appraisal processes that are organized by this Center derive directly from the number of participants enrolled in its appraisal processes. Based on these assumptions, this work presents a model for the use of machine learning techniques on the logistical organization of the evaluation processes organized by Comperve. The model presented here was created from the data bases available in the Center, which contained information about the evaluation processes carried out by Comperve since the beginning of the 2000s. In order to carry out this work, the execution context of the activities where this Center is currently located was investigated, analyzing how the logistic organization of its processes is done, integrating the data that was decentralized and de-standardized, and creating the training model that achieved more than 98% of accuracy in the classification of the quantity of participants enrolled in its processes. For the application of this model, an application using the infrastructure of the management system of the appraisal processes, which is currently being developed, was developed.