Analyze the Impact of Emerging Terms in Semantic Similarity for Effective Data Mining Applications

Abstract

Due to the emerging web searching process and the requirement of newlinefast and efficient result provisioning, the semantic similarity becomes a newlinesignificant objective in Information Retrieval (IR). Use of static lexical newlineresources leads many conventional semantic similarity measurement newlinetechniques to ignore temporal aspects of concepts, resulting in an newlineinconsistent semantic measure and IR. In reality, the semantic processing newlineor data mining technique is applied to automatically extract the relevant newlineinformation from lexical ontologies that contains the list of semantically newlinerelated terms of a particular term. This research work delivers two newlinesignificant contributions in evolving terms that impact the applications newlinethat depend on the semantic similarity measure. newlineThe first contribution of this work exploits the sources of disease newlineconcepts captured across biomedical resources and automatically newlineidentifies the medical terms to extend the Disease Ontology. The newlineExtending disease ontology with newly evaluated terms to improve newlinesemantic medical information retrieval (ENHANCE) system associates the newlineextracted terms with its correct ontological class to extend the Disease newlineOntology. The process of vectorization determines which anatomy newlineontology content represented by the biomedical resource has to be newlineenriched using dynamic n-gram mapping and linkage distance metric. newlineThen, the ENHANCE extracts the disease name with its relative symptoms newlineby removing the non-contextual information using the Word Net and newlineDisease Ontology. The experimental results prove that the ENHANCE newlinesystem provides an improved biomedical document retrieval system in newlineterms of achieving 11.49% precision value while submitting the new newlinedisease terms than the baseline sibling discovery method. newlineThe second research utilizes the social media as the robust, dynamic newlinesource in providing the precise result for recent trend concepts. This work newlineintroduces an identifying new concept Evolution through semantic Social newlinemedia temporal analysis (NEOTERIC-SPY) model. Initially, the NEOTERIC newlineSPY measures the semantic relatedness of a word-pair using normalized newlinedistance measurement in a time series. It reduces the word-pair list of a newlineconcept among the overall possible word-pairs. Then, it identifies the new newlineconcepts using temporal correlation analysis of the semantic similarity newlinemeasurement between the time series. The experimental results show newlinethat the NEOTERIC-SPY approach significantly outperforms the newlineconventional method and illustrates the consistent improvements in its newlinerecall by 9.78% than the baseline ESA method. newline newline newline