In order to develop manpower expertise in the area of oil, natural gas and biofuels, the National Petroleum Agency - NPA - has been promoting training programs for Human Resources (PHR). Aimed at determining the strategies behind these programs in a structured and comprehensive manner, methods of administration such as Balanced Scorecard (BSC) have been adopted. However, failures such as non-measurement of uncertainties involved up until the desired goals have been reached, suggest the adoption of more sophisticated formulae, such as Bayesian networks (BN) rather than BSC. In the case of BN, the challenge becomes a proper quantification of its parameters, allowing reliable estimates from them. The objective of this work is to evaluate the performance of BN in handling a given real PHR problem by means of the comparison between its predictions and the respective observed results. The model has been developed from a BSC and subsequently quantified according to expert opinions, since no relevant empirical data was available. The BN was designed to support the deployment of strategies for scientific development and opening of markets for a given university course. The analyses carried out indicate the good performance of the model, emphasizing the adopted elicitation method in such a process.
Program for human resources; Strategic planning; Balanced scorecard; Bayesian Networks; Methods of elicitation; Collection of opinions
