Agent Based Multilingual Information Retrieval System: A Design Approach

Author Name(s): Mangala Madankar, Manoj Chandak, Dr. U. N. Shrawankar
Author Email:


As India is a Multilingual Country and on internet multilingual database content increases day by day. Hence multilingual information retrieval is very curious in a country like India. In this paper, a new research design of agent based multilingual information retrieval system is proposed. Agent based architecture is designed for the system architecture development. Here system will support the queries in Hindi, Marathi and English language. This paper briefly describes the multilingual information retrieval design and development system and the role of various agents in IR system.



The utilization of web surfing is developing now daily, different data necessities of clients in dialects unique from English are expanding quickly. The quickly quickening pattern of globalization of the achievement of e-Governance arrangements and organizations expect information to be put away and recovered in a wide range of regular dialects. The essential information stockroom for such applications, should be proficient as for multilingual information. Productive multilingual information stockpiling and inquiry preparing of information straddling over in excess of one regular dialects are of vital significance in the present globalize world. CLIR manages asking inquiries in a single dialect and interest for recovering archives in another dialect. MLIR tolerating that making inquiry in various dialect and recovering archives in various (more than) the question dialects. Our framework contain immense measure of information and information can be controlled in the database by a few counterfeit dialects. Numerous Natural Language frontends have been created as a win Work created in the region of artificial intelligence, yet the vast majority of them utilize the English as a characteristic dialect. India being a Multilingual Nation and just 5% of the population brag the training up-to matriculation level and there utilize is constrained. The information retrieval in local languages from storage database has majority of the impact. [1] This work proposed a framework
of an agent based architecture that utilizes the various intelligent agents in the multilingual IR system. In the proposed research design multilingual database like resource information in Hindi, Marathi and English language of an education and sport domain will be stored in cloud environment. The user can give his query to the system in his native language with the help of user interface agent. Analyzer agent plays an important role for the multilingual query processing. User can ask the query in any language and information stored in database will be of any language from Hindi, English and Marathi, such as user can ask the query in Marathi and access the information in Hindi. The agents will be written in the programming language and they will communicate via Inter-agent Communication Language (ICL). Various agents will be involved here such as User interface agent, Database agent, Analyzer agent, Server agent, Parser agent and SQL agent.


Cross-lingual and Multi-lingual Information Retrieval System gives new pattern to seek records through various assortments of dialects over the world and this can be the new standards for looking among two dialects as well as in numerous. User can enter the query in Marathi language or the Hindi language but in the database information are stored in the English language, and then our agent based system will convert the Hindi or Marathi query into English and find the related information from the database and again convert the document into the user native language like Hindi or Marathi. We are used the precision and recall methods to find the exact relevant information in the users native language. Experimental set up is ready, Cloud design and interface modules are ready; results will be shown in next version of the paper.

370 total views, 1 views today

Download PDF File

About the author: admin