63ª Reunião Anual da SBPC

A. Ciências Exatas e da Terra - 2. Ciência da Computação - 2. Banco de Dados

A WEB INFORMATION RETRIEVAL SUPPORT SYSTEM IN PATIENT INFORMATION LEAFLET

Sandra Regina Rocha Silva 1 Cristiano Expedito 2 Fábio Galluzzo 2 Adriana Vivacqua 2 Moacir Florentino da Silva Junior 3 Geraldo Bonorino Xexéo 4

1. Programa de Pós-Graduação da Universidade Federal do Rio de Janeiro - COPPE/UFRJ 2. Programa de Pós-Graduação em Informática Universidade Federal do Rio de Janeiro - UFRJ 3. Instituto Federal do Maranhão - IFMA 4. Prof. Dr./Orientador - Programa de Pós-Graduação da Universidade Federal do Rio de Janeiro - COPPE/UFRJ

INTRODUÇÃO:

Web information retrieval currently faces major challenges such as the problem of information explosion, user difficulty in establishing one’s search, and the difficulty in managing the complexity of the search process. Some Web information retrieval support systems (WIRSS) are targeted at specific domains such as the patient information leaflet domain. The existing WIRSS for patient information leaflets are complex to use, as their search models still do not follow the latest WIRSS trends. We propose a WIRSS model in patient information leaflet that follows the new trends in the area. The model envisages the use of query expansion, faceted search, mind maps, and user profiles.

METODOLOGIA:

Initially we raised the main WIRSS in patient information leaflet and evaluated their search models. Then we surveyed the following new key trends in WIRSS: helping users to specify their search requirements, evaluating and exploring search results, and managing the information returned. We found that current WIRSS in patient information leaflet did not follow these WIRSS trends. We examined the current literature in the subject, looking for possible proposals to meet current WIRSS requirements. We then selected, amongst the proposals we found the most appropriate one to build our WIRSS model in patient information leaflet. We implemented a WIRSS prototype in patient information leaflet based on our model. The prototype used 900MB base, corresponding to 3,843 pdf documents as information leaflet corpus, downloaded from http://www.onofre.com.br/onofre/bulas.aspx. The construction of the prototype used the following tools: Lucene, to index and search the base; JSF1.2 (Java Server Faces) to build the interface for user interaction; SVN Google code to manage application versions; Apache Tomcat as the Java Servlet; and a Java Script application programming interface to construct mind maps.

RESULTADOS:

We propose the use of query expansion to help users in building their queries. Query Expansion is the term given when a search engine adds search terms to a user’s weighted search. The domain of patient information leaflet has a specific terminology, and we propose the process of automatic query expansion using a computer Thesaurus to help the user build a valid query. Our model proposes the use of faceted searches to explore the results returned. Faceted searching is a technique to access a collection of data items represented, using a faceted classification, allowing users to explore by filtering available information. A faceted classification system allows the assignment of multiple classifications to an object, enabling the classifications to be ordered in multiple ways, rather than in a single, pre-determined, taxonomic order. Each facet typically corresponds to the possible values of a property that is common to a set of digital objects. Our model proposes the use of information visualization and the definition of user profiles to manage the information returned. Information visualization was implemented by constructing a mind map for the search results. The user is at the centre of the mind map and one’s profile is used to filter only data that is useful to him/her.

CONCLUSÃO:

Our proposal entails the use of hyperlinks to organize retrieval results and also allows access to the full contents of the leaflet. Furthermore, we propose visualizing search result through the construction of a mind map associated to the user profile. The user appears at the centre of the map and the drugs are associated with him/her on the map. Thus the information that appears on the mind map is directed to the profile of this user. For an user aged one, the search result on the map visually indicates whether a particular medicine is suitable for this user. If the drug is indicated for that user, the map shows information suitable for the profile of this user, such as, for example, only the dosage recommended for his age group. The model proposes the use of query expansion to help an user build one’s query, the use of faceted searches to explore the results returned, the use of mind maps, and the definition of user profiles to manage the information returned. The proposed model is innovative and follows the latest trends in WIRSS. We implemented a first prototype for the proposed model. The prototype is currently being refined.

Palavras-chave: Faceted search, Information visualization, Query expansion.