Using Bayesian networks to discover relationships between bibliometric indices. A case study of computer science and artificial intelligence journals

As they are used to evaluate the importance of research at different levels by funding agencies and promotion committees, bibliometric indices have received a lot of attention from the scientific community over the last few years. Many bibliometric indices have been developed in order to take into account aspects not previously covered. The result is that, nowadays, the scientific community faces the challenge of selecting which of this pool of indices meets the required quality standards. In view of the vast number of bibliometric indices, it is necessary to analyze how they relate to each other (irrelevant, dependent and so on). Our main purpose is to learn a Bayesian network model from data to analyze the relationships among bibliometric indices. The induced Bayesian network is then used to discover probabilistic conditional (in)dependencies among the indices and, also for probabilistic reasoning. We also run a case study of 14 well-known bibliometric indices on computer science and artificial intelligence journals.