Information Retrieval through Multi-Agent System - Research Paper Example

Research Paper Information Retrieval through Multi-Agent System with Data Mining in Cloud Computing Introduction Problem solving solutions like Multi-Agent System (MAS) capitalizes on its multiple intelligent agents to receive precepts from the environment, process the information and produce the desired result for the environment. The wide scope of capabilities of an MAS permits the user to resolve functional, methodic, algorithmic or procedural query to explore and process the data. MAS are also referred as autonomous agents having the capability to resolve problems that are not possible for a single agent to handle. The aim of this research paper is to develop a practically implemented research model for the information retrieval using Multi-Agent System with Data Mining technique in a Cloud Computing environment. The paper will undertake a review of the existing literature available on this arena and develop an empirical model showing real time data flow through MAS with data mining after retrieval of meaningful information from data warehouse present in a cloud computing environment. In the end, paper will provide recommendations for the organizations for effective implementation and use. Definitions Cloud Computing is a general term that refers to anything that “involves delivering hosted services over the Internet. Broadly it is characterized into three categories, namely: Software-as-a-Service (SaaS), Infrastructure-as-a-Service (IaaS) and Platform-as-a-Service (PaaS)” (‘CloudComputing’, 2007). Multi-Agent System is a problem solving “system composed of multiple interacting intelligent agents” (Kent, Kobti, Snowdon and Aggarwal, 2010). Data Mining means “Discovery” of new knowledge that was not known before. It is “the analysis step in the Knowledge Discovery and Databases process” (Nodine, Ngu, Cassandra and Bohrer, 2003). Data Warehouse refers to a data store which involves three stages, namely: staging, integration and access for reporting and analysis purposes. Related Works In the present age and knowledge economy, discovering new knowledge and retrieving information from a data center from a cloud environment is a difficult aspect. The concept of cloud computing does not provide facilities for the knowledge discovery and information retrieval. Furthermore, it is required that the so-called knowledge discovery should be in harmony with the structure, schema and architecture of that knowledge. The emerging knowledge cloud is considered insufficient to retrieve information effectively and thus, Chang, Yang and Luo (2011) undertook a study to propose "an ontology-based agent generation framework for information retrieval in a flexible, transparent and easy way on cloud environment” (p.1135). They proposed a framework for information retrieval in which the user will submit "a flat-test based” request to retrieve “information on a cloud environment”, the request will be “deduced by a Reasoning Agent automatically that is according to a predefined ontology and a reasoning rule and then translated to a Mobile Information Retrieving Agent Description File (MIRADF) that is formed in a proposed Mobile Agent Description Language (MADF)" (Chang, Yang & Luo, 2011; p.1136). MIRA-GA, a generating agent will generate MIRA in accordance with MIRADF. Nodine, Ngu, Cassandra and Bohrer (2003) studied InfoSleuthTM system that is a unique discovers and retrieves information in real time and open environments. Primarily, this system is based on MAS. InfoSleuthTM has a multi-broker function which interacts with syntax and semantics in MAS domain. A broker has to decide through the advertised information regarding the qualities of each agent, which agent is best for a particular service. Thus, InfoSleuthTM provides a brokering system where brokers can share and receive information regarding other brokers and non-brokers to decide with which broker to finalize the deal. Lai, Chang, Hu, Huang and Chao (2011) used data mining technique by analyzing large-scale data and obtaining statistical answers from it regarding suggestions for TV programs. The study weighed these figures through various techniques of data mining regarding preferences of viewers of TV programs in a particular area. The authors found that this process involved large amount of data, computer power problems and scale as the major bottlenecks in the effective implementation of the techniques in finding the most viewed TV programs. They suggested an architecture involving cloud computing and "a map-reduced framework, map-reduce version of k-means and the k-nearest neighbor (kNN) algorithm is applied" (Lai, Chang, Hu, Huang and Chao 2011; p. 124). Literature Review The contemporary research available on the concept of cloud computing, multi-agent system, data mining and data warehouse revolves around the identification of bottlenecks in existing data flow frameworks and introducing new frameworks that overcome these problems. The extensive data available regarding these concepts highlight the wide scope of cloud computing and multi-agent systems when used with data mining techniques for information retrieval. Manvi and Venkataran (2004) postulated that software agents, sub category of MAS, are popularly used communications specifically the mobile agent technology. The authors found that MAS has flexible operations, easy to use, adaptable to real environments and wide reliance on the concepts of software engineering. InfoSleuthTM and other relative systems had issues regarding Agents-Communication Language (ACL) and other knowledge domains (Nodine & Chandrasekara, 1999). These systems lacked the capabilities to satisfy the emerging needs of the environment with the expansion of the scope of MAS application and increase in its complexity. The existing MAS based information discovery and retrieval systems became insufficient in terms of information security, information transfer and the usage of information on multimedia. Kent, Kobti, Snowdon and Aggarwal (2010) studied unified data management and decision support system (UDMDSS) in the light of health care. They developed a system that is founded on a modular architecture that supports semantic data models, queries, Bayesian statistical analysis, data acquisition through mobile, artificial intelligence for simulation and modelling in health care and web-base desktop in real time environments apart from various other features. They focused on the Canadian hospital and children safety from car accidents. Multi-Agent Systems are autonomous at least partially. No one intelligent agent has a complete view of the entire system due to complexity of the system that limits the use of knowledge completely by a single agent. Furthermore, there is no defined control system that forms it a monolithic system. Data mining on the other hand, automatically or semi-automatically studies large number of data to find similar patterns that are previously unknown. In most cases the patterns found by data mining technique are processed through decision support systems to further analyze the data for multiple purposes and uses. Cloud computing is centralized that makes it less costly, convenient, high utility, scalability, agility, reliability, performance and security of data. There are five layers in cloud computing which are equally effective to share information once the Internet has established connection with two or more computers, namely: client, application, platform, infrastructure and server. There is a public, community, hybrid and a private cloud which have advantages as well as disadvantages. The integrated data warehouses integrate and collect data from various areas of the business at one point for the users to analyze them and benefit from the required information of choice. Methodology The research paper adopted a qualitative research method based on the secondary data collection activities. The preliminary use of knowledge available on the cloud computing, data mining, data warehouses and multi-agent system assisted in the formation of a new research model with practical significance and utility. The peer-reviewed journal articles, books and periodicals provided a concrete foundation for the invention of a new model for information retrieval. This justifies the adoption of qualitative research method using secondary data collection technique. Findings and Summary The study of the related works and literature review that in order to retrieve meaningful information from the data warehouse through the help of a multi-agent system and data mining techniques in a cloud computing environment, the following architecture is designed: The Infrastructure-as-a-Service (IaaS) provides a virtual environment with storage and network without having physical hardware. Therefore, Infrastructure cloud computing provides a data warehouse for storage of data for further analysis. The user will submit a flat-text based request on the IaaS for information retrieval from its integrated data warehouse that has gathered data from numerous areas of business to present to the user a wide variety to choose from. The IaaS will forward this request to the MAS to find the information as requested. However, MAS does not have the ability to find the information that has large amounts of data from a data warehouse. Therefore, it will use the data mining algorithm to analyze the large amount of data from the data warehouse that is residing in the IaaS. As a pre-processing stage, the MAS will first develop a target data set which will be large enough to contain all the possible data patterns and send it into the system. Then the processing will begin where the data will be analyzed through anomaly detection, clustering, classification, regression and summarization. The result of the process will be shown on the screen to the user. Figure 1: Information Retrieval through Multi-Agent System with Data Mining in Cloud Computing Conclusion and Recommendations The information retrieval practical model through the multi-agent system with data mining in a cloud computing environment has been proposed. It is however, recommended that users should ensure that the request made to the IaaS is within the scope of integrated data warehouse and is clear and simple. Thus, making the work for the multi-agent system easier through application of the data mining algorithms to retrieve meaningful information from the data warehouse. In this proposed research model/architecture, the use of cloud computing allows the users to retrieve meaningful information from virtually integrated data warehouse that reduces the costs of infrastructure and storage. References Chang, Y-S., Yang, C-T, & Luo, Y-C., (2011). An Ontology based Agent Generation for Information Retrieval on Cloud Environment. Journal of Universal Computer Science, Vol. 17, No. 8, Pages: 1135-1160. Retrieved October 25, 2011 from http://jucs.org/jucs_17_8/an_ontology_based_agent/jucs_17_08_1135_1160_chang.pdf ‘Cloud Computing’, (2007). Cloud Computing Definition. Retrieved 25 October 2011 from http://searchcloudcomputing.techtarget.com/definition/cloud-computing Kent, R.D., Kobti, Z., Snowdon, A., & Aggarwal, A., (2010). Towards a unified data management and decision support system for health care. Intelligent Interactive Multimedia Systems and Services, Vol. 6, pp. 205-220. Retrieved October 25, 2011 from http://www.springerlink.com/content/u318387487131428/ Lai, C-F., Chang, J-H., Hu, C-C., Huang, Y-M., & Chao, H-C., (2011). CPRS: A cloud-based program recommendation system for digital TV platforms. Journal Future Generation Computer Systems, Vol. 27, Issue 6. Retrieved October 25, 2011 from http://dl.acm.org/citation.cfm?id=1967928 Manvi, S.S., & Venkataram, P., (2004). Applications of agent technology in communications: a review. Computer Communications. Vol. 27, Issue 15, Pages: 1493-1508. Retrieved October 25, 2011 from http://www.sciencedirect.com/science/article/pii/S0140366404001914 Nodine, M.M, Ngu, A.H., Cassandra A., & Bohrer, W.G., (2003). Scalable semantic brokering over dynamic heterogeneous data sources in InfoSleuth™. Knowledge and Data Engineering, IEEE Transactions. Retrieved October 25, 2011 from http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?arnumber=1232266 Nodine, M., & Chandrasekara, D., (1999). Agent Communication Languages for information-centric agent communities. Systems Sciences, Pages. 1-10. Retrieved 25 October 2011 from http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?arnumber=773094 Read More