The Information Retrieval Process of the Scientific Production at Departmental-Level of Universities: Exploration of New Approach

Introduction Our work was focused on document retrieval from Scopus databases of the Escuela de Ingeniería de Sistemas y Computación (EISC) of the Universidad del Valle (Cali Colombia). The databases systems as WoS (Web of Science) or Scopus contain the knowledge produced by engineer schools. However, this information is ambiguous and the retrieving of the specific documents of one school is identity uncertainly (Pasula et al., 2003). Thus, the design of machines (search engines) to retrieve the relevant documents of engineer schools is a complex process. After the work of Bucheli et al. (2013); Cuxac, Lamirel, & Bonvallot (2013) proposed a semisupervised approach, mixing soft-clustering and Bayesian learning. Additionally, Huang et al. (2014) proposed a rule-based algorithm. Both approaches were for affiliation disambiguation. We reproduced the model proposed by Bucheli et al. (2013). The results show that the model can be used to information retrieval of department-level. In addition, we proposed a new approach addressing the problem of classification using network science. The future work will be related with building a model according to the network science approach.