A novel fuzzy result ranking technique

Database queries usually consist of elementary query conditions combined with each other using connectives. Elementary query conditions are typically Boolean-valued functions defined over one or more properties. These elementary query conditions may be described in an imperfect way, or they may be described perfectly, but correspond to attributes for which the database values might be subject to imperfection. Whenever one of these cases appears, the evaluation of the elementary query condition for the database value is often quantified using one or more numbers called ‘quantifications’. Examples of quantifications are satisfaction degrees, possibility degrees, chances etc. Ideally, these quantifications would be combined somehow to allow ranking the query results in a way that reflects the extent to which they comply with the user's query. However, this poses some major challenges. For example, the way in which quantifications of different types may be combined with one another is not clear and seems unstraightforward, as does the way in which the query results must be ranked or ordered to present to the user. The work presented in this paper attempts to deal with these challenges by proposing a novel technique for quantification aggregation and query result ranking. Using this novel aggregation technique, some extra information is used to transpose quantifications of different types to quantifications of a single type, completely bypassing the aforementioned issues.