Information retrieval in multilingual environment phonetic and semantic along with ontology approaches for hindi and marathi

Abstract

Information retrieval (IR) is an important issue and difficult to handle especially in newlinemultilingual environment. The IR process begins with entering a query into the system in newlineprescribed format. A query does not uniquely identify a single object but several objects with newlinedifferent degrees of relevancy. An object is an entity from database which represents the newlineinformation. User queries are matched against the database information. The matched objects newlineare ranked according to the query terms and shown to the user as information. newlineMultilingual environment (ME) provides a platform where a query can be inputted in one newlinelanguage and the result can be in the same language or different language. This result can be newlineviewed as IR or knowledge. Further this IR helps to make better decision in various newlinescenarios. This is only possible in a system that supports multilingual environment in all newlineaspects. In order to enhance the IR process of a system, phonetic and semantic issues have to newlinebe resolved. The phonetic and semantic matching will enhance the efficiency and utilization newlineof resources for IR. International Phonetic Association (IPA) as phonetic standard gives a set newlineof alphabet as a standard way of representing any language string that is utilized as phonetic newlinematching. Ontology is a promising solution to any semantic matching problem. It typically newlineprovides a vocabulary that describes a domain of interest and specification of meaning of newlineterms used in the vocabulary. newlineSeveral different solutions have been proposed for phonetic and semantic matching so far. newlinePhonetic matching solutions worked either on developing an operator in a system or newlinedesigning rules as per phonemes of alphabets of a language. The operator uses text-tophonetic newline(TTP) system for translation of a text string into its equivalent phonetic form. This newlineform represents the alphabets from IPA. But TTPs for Hindi or Marathi are cumbersome to newlineuse. In semantic matching, solutions proposed so far were based on ontology, where some newlinesystem used top-down ap