Vertical searching in juridical digital libraries

In the world of modern digital libraries, the searching for juridical information of interest is a current and relevant problem. We approach this problem from the perspective that a new searching mechanism, specialized in the juridical area, will work better than standard solutions. We propose a specialized (or vertical) searching mechanism that combines information from a juridical thesaurus with information generated by a standard searching mechanism (the classic vector space model), using the framework of a Bayesian belief network. Our vertical searching mechanism is evaluated using a reference collection of 552,573 documents. The results show improvements in retrieval performance, suggesting that the study and development of vertical searching mechanisms is a promising research direction.