Search Engines to QAS: Explorative Analysis
Sampat P. Medhane, Shashank Joshi, Snehal Shinde · 2014
Search engine is pedestal machine on two-dimensional logic and possibility hypothesis which are short of concept based i.e. fact based existent human understanding and advanced way of thinking and precision to formulate answers.QA (Question answering) system is hybrid machine with analysis ability to fetch data from diverse knowledge source assorted format and formulate a answer to user search query. Research domain in advanced Information retrieval is machine to prospect a information seeking engine (search engine advancement).formulate precise answer to user search question is aspire of QA system .current search engines are keyword base with sort of semantic incorporation (semantic search in Google) Enhancement in search engines is three tier, user question understanding (user model, context, intent), text/data-mining analytics with machine learning (Support vector machine(SVM),Artificial neural networks(ANN)),and adha Rank (bings algorithm) to formulate trainable machine to reduce search space and time.QA system is a semantic concept based framework which identifies the nucleus keywords and phrases incorporating meaning i.e. sense of word to phrase ,sense of phrase to sentence ,to corpus which abolish the disambiguation of NL(natural language) in search engines. A prototype design of Answer engine is proposed that performs concept map of user question .the semantic concept framework enhances the ability of reckoning in Answering system. GUI interface performs mapping of user question in concept group (ontology class i.e. Spatial ontology, domain ontology, task ontology, universal ontology).text analytics formulates extraction of Answer term, phrases from corpus or knowledge dataset (unstructured & structured) using n-gram. Machine learning algorithms facilitate intelligence in search corpus. Ranking algorithm presents information on relevance precision mapped to user search intent. AdaRank repetitively build ‘frail rankers’ on the foundation of re-weighted training data and lastly linearly join the feeble rankers for making ranking forecast. The training process of AdaRank is exactly enhancing the performance measure used. AdaRank significantly optimizes the base technologies BM25, Ranking SVM, and Rank Boost.QA system are future of search engine, expert systems, and rule based systems. Current paradigm information processing machines to intelligent information machine which perform time variant search on large dataset with higher precision to question. This paper a search analysis has been carried out on 20 research article on Advance IR technologies. The search analysis present in detailed study on three methodologies (question understanding, text analytics, machine learning and ranking algorithm) when optimized invent enhanced search machine, QA system. Search paper concludes that intelligence in information processing has evolved search engines, rule based machines, expert systems to hybrid reasoning machines-QA Systems.