Research Of Personalized Intelligent Information Retrieval Based On Multi-Agents

Information retrieval technology first appeared in the library of dataretrieval system, along with the development of computer science,information retrieval areas continue to develop and grow, and it goes beyondthe standard citation and a set of useful literature retrieval from the initialgoals. Retrieval of information has been included in the areas of modeling,and Literature Classification classified, system builders, user interface, datavisualization, and information filtering and query language. The rapiddevelopmentofthe Internet,informationretrieval intoanewstage,hadaWebinformation retrieval of the most widelyused method, the Web search engine,its appearance to some extent solve the network users access to informationfrom the difficulties. However, applications of various information retrievalsystems are unable to provide personalized information retrieval services,personalized intelligent information retrieval technology as the current hotspotsofstudy.General information retrieval techniques based on keywords to usesimple matching retrieval mode, todayis timelyaccess to the huge success oftheapplicationofinformationretrieval searchengineis stillintheuseofsuchretrieval methods. It will surely result of the current information retrievalsystemhavesomeshortcomings,mainlythefollowing:1. Lack of effective means of user interaction, not timely and effectiveaccess to user feedback, it can only passively accepted user enquiries, and donothaveactiveserviceintheuser'sinformationdemandcapabilities;2. Different interest with their clients did not provide personalizedinformation retrieval services, systems consider only queries and indexingdocuments matchingthe word,regardless ofwho theusers are, andregardlessofuserinformationtothoseinterested;3.AccordingtotheKeywords matching,thesearchresults often containa lot of irrelevant information; users will need to spend considerable time andeffortinseekingtoreturntotheoutcomeofinterestingresults.How to Design and Implementation of an effective remedy General Information Retrieval System personalized intelligent retrieval system, is acomplex issue. First, we must realize we must change the personality of theoriginal models for keyword matching, because keywords can not be fullyrepresentative of a user's personality and their specific information needs, theneed for a description of other user information needs, this Interested userscan be adopted to solve the model for users interested in a model that thepersonality characteristics of user information, a user interest model,personalized information retrieval system can use its information filtering,users will be really interested Back to the user. Second, we must realizepersonalized smart Intelligent Information Retrieval System, a system toallow users interested in learning ability and can take the initiative to thecompletion of the mission objectives. Agent is the product of artificialintelligence field, it is a class in particular circumstances can sense theenvironment, and to achieve a series of independent objectives of thecalculation procedures, or entities, its appearance to address the complexbehavior characteristics ofthesystem provides a powerful modelingThetool,it is natural that Agent technologycan be applied to Personalized InformationRetrieval system. At present personalized search technology in the moremature technology content-based filtering methods and collaborative filteringmethods, these methods require the user to filter model of interest, so usersinterested in modeling is an important personality of the search area, whilesomesystemsAgentalso usethetechnology,canbepersonalizedsearchallofthe different functional decomposition of the Agent to achieve in order toobtainbetterperformance.In order to solve the existing information retrieval technology existingproblems through relevant, personalized information retrieval system toachieve the objectives of the research function, as well as Agent Technologyanalysis, which is based on user interaction Agent, users interested in modelAgent, enquiries Results and data filtering Agent Information ProcessingAgent four Personalized Information Retrieval System Design andImplementation, through mutual cooperation between them in order toachieve personal Intelligent Information Retrieval System. AlthoughAgent-oriented programming language and environment are not maturesupport, but the maturity of the existing object-oriented technology and the realization of the Agent provide some basic support. The system employs aexpansionofexistingmatureobject-orientedtechnologyandmeanstoachievethevariousAgent,anumberofmethodstoachievethistypeofobject-orientedsoftware system to achieve by the Agent should have a function, and caneasilyImplementationoftheframeworkwillbeappliedtoeachspecificAgentDesignand Implementation,eachspecificAgentachievetheirspecificonlytoconsider the information of interest, the corresponding related events as wellas its implementation of the plan can be, so the whole system the design of aclear structure, it will be convenient for everyone Agent needed to be revisedorexpanded,makingagoodsystemmaintainability.As a personalized intelligent information retrieval system a key part, theauthor focuses on the user modeling and information filtering results of thevarious methods and the main algorithm, and gives concrete realization.Experiments show that, the paper is based on Multi-Agent TechnologyModeling and users interested in information filtering technologyPersonalized Information Retrieval, the average absolute error is only0.1056723,andachievedmoresatisfactoryresults.The work done by the author is only based on the personality of theIntelligent Information Retrieval System Design and Implementation of anumber of exploration and try, there are still many issues need to be furtheraddressed.Theauthor'sfurtherworkneedstobedoneinclude:1. Modeling algorithm for more tests, compared on the basis of themodeling algorithm to improve or choose a more suitable modelingalgorithm;2. Part of the communications system to improve and enhance thecollaborationcapacitybetweentheAgents;3. Use more evaluation methods and data to test the system, a bettersystem to measure the performance level, and find the key points can beimproved.