Resumo

RUIZ LOBAINA, Esther Marina  e  ROMERO SUAREZ, Pedro Lázaro. Finding patterns to improve products and services in libraries. Investig. bibl [online]. 2017, vol.31, n.72, pp.209-225. ISSN 2448-8321.  https://doi.org/10.22201/iibi.0187358xp.2017.72.57830.

This work shows the results obtained during the search for hidden patterns using statistical algorithms, in a bibliographic database. For this research the WinIDAMS v.1.3 software, used for data management, building an IDAMS dataset (BUILD) and Data Group (AGGREG) for statistical analysis algorithms, Cluster analysis was selected (CLUSFIND) and scatterplots (SCAT). In addition to the outputs of the results this software provides multidimensional tables, able to create for each group of selected variables, an internal table with results such as frequency and average arithmetic were selected for these tests, while for graphical representation of the results, it was decided to use histograms, because they are bar graphs that allow us to interpret very easily and quickly, the behavior of the variables selected for analysis. This study found patterns through clustering, with which services could enhance Selective Dissemination of Information and propose new services, to become part of the products offered by the library.

Palavras-chave : Information Science; Information Management; Multidimensional Tables; Data Mining; Statistics; Patterns; Histograms.