|
Abstract : |
By using data abstraction concepts from database and objectoriented systems and combining them with uncertain inference, we develop a new approach for the design of information retrieval (IR) systems, thus solving major problems with regard to networked IR. Different data types with vague predicates are required to allow for queries referring to arbitrary attributes of documents. Physical data independence helps in solving text search problems related to noun phrases, compound words and proper nouns. Logical data independence allows for different views on an IR database, thus implementing e.g. data security. Inheritance, especially on attributes, supports the creation of unified views on a set of IR databases. Uncertain inference allows for query processing even on incompatible database schemas. 1, |
