In most commercial online systems, the retrieval system
Of course not all commercial services are based on the is based on the Boolean model and its inverted file orga-Boolean schemes, and the WIN retrieval system is widely nization. Since the investment in these systems is so considered as an exception to this rule (Turtle, 1994).
great and changing them could be economically unfeasi-As an alternate retrieval scheme, the vector-processing ble, this article suggests a new ranking scheme espeor the probabilistic retrieval models use both weighted cially adapted for hypertext environments in order to produce more effective retrieval results and yet maintain document and query vectors. In this case, the size of the the effectiveness of the investment made to date in the output is kept under control using a threshold for query-Boolean model. To select the retrieved documents, the document similarity or probability of relevance. Unfortusuggested ranking strategy uses multiple sources of nately, these models require more processing time and document content evidence. The proposed scheme intethey do not allow structured queries (synonym, termgrates both the information provided by the index and query terms, and the inherent relationships between phrase).
documents such as bibliographic references or hyper-
In this article, we consider request structure and comtext links. We will demonstrate that our scheme repreputational efficiency to be two important facets of a comsents an integration of both subject and citation inmercial information retrieval system. Moreover, our indexing, and results in a significant improvement over classical ranking schemes used in hybrid Boolean systent is to suggest a new ranking scheme which preserves tems, while preserving its efficiency. Moreover, through the large amount of investment made in the Boolean knowing the nearest neighbor and the hypertext links model. We also prefer proposing a simple but general which constitute additional sources of evidence, our technique that can be applied to most existing systems strategy will take them into account in order to further improve retrieval effectiveness and to provide ''good'' instead of a specialized solution applicable only in narrow starting points for browsing in a hypertext or hypermedia contexts. In this vein, our study presents a new ranking environment.