Information retrieval per la navigazione assistita

The aid that an user can get looking up a collection of documents is tightly correlated to the quality of provided informations. For this reason the research step should not be seen as the automatic creation of a ranked list of documents from a query, but it should also attend the user in his navigation process, providing resources as near as most to his information need.

In this work we present a system that integrates the requested document model with the user model. This type of integration is made in the navigation step, combining the dynamic query with the requested static document, enriching it with links to the most relevant resources for the user.

As case of study we choose the Python documentation. This type of resource, available on the web, shows hyperlinks inserted by the authors - or by tools that they use - that connect documents using their hierarchical structure and using a set of static links, that however aren't suitable for the user. The goal is reached by the creation of an adaptive interface that attends the user in his navigation process, inserting links to related documents and permitting him to reach the needed informations, in our opinion, more simply compared with the traditional navigation. The system, moreover, can integrate more sources, sometimes heterogeneous, that provide additional informations to the single document, enriching its navigation paths and, consequently, its information need.