Maheswari, J.U.; Karpagam, G.R.: ¬A conceptual framework for ontology based information retrieval (2010)
0.00
0.00394768 = product of:
0.027633758 = sum of:
0.027633758 = weight(_text_:ed in 702) [ClassicSimilarity], result of:
0.027633758 = score(doc=702,freq=2.0), product of:
0.140671 = queryWeight, product of:
3.5559888 = idf(docFreq=3431, maxDocs=44218)
0.0395589 = queryNorm
0.19644247 = fieldWeight in 702, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
3.5559888 = idf(docFreq=3431, maxDocs=44218)
0.0390625 = fieldNorm(doc=702)
0.14285715 = coord(1/7)
- Abstract
- Improving Information retrieval by employing the use of ontologies to overcome the limitations of syntactic search has been one of the inspirations since its emergence. This paper proposes a conceptual framework to exploit ontology based Information retrieval. This framework constitutes of five phases namely Query parsing, word stemming, ontology matching, weight assignment, ranking and Information retrieval. In the first phase, the user query is parsed into sequence of words. The parsed contents are curtailed to identify the significant word by ignoring superfluous terms such as "to", "is","ed", "about" and the like in the stemming phase. The objective of the stemming phase is to throttle feature descriptors to root words, which in turn will increase efficiency; this reduces the time consumed for searching the superfluous terms, which may not significantly influence the effectiveness of the retrieval process. In the third phase ontology matching is carried out by matching the parsed words with the relevant terms in the existing ontology. If the ontology does not exist, it is recommended to generate the required ontology. In the fourth phase the weights are assigned based on the distance between the stemmed words and the terms in the ontology uses improved matchmaking algorithm. The range of weights varies from 0 to 1 based on the level of distance in the ontology (superclass-subclass). The aggregate weights are calculated for the all the combination of stemmed words. The combination with the highest score is ranked as the best and the corresponding information is retrieved. The conceptual workflow is illustrated with an e-governance case study Academic Information System.
