Data Mining and Integration of Real-time Sensory Information in a Cloud Framework

Abstract

In present times data is a ubiquitous resource, and its analysis to deduce highly selective and relatively more useful bits of information often holds great signi cance. However, in most advanced systems, the deduction of useful insights from a large dataset denotes only part of the required functionality; because the information derived from the input data needs to be integrated back into the system to increase the overall e ciency of operations in terms of feedback-based predictive capabilities or intelligent decision making. Preferably, this integration of information in a dynamic environment needs to be carried out autonomously with minimal human intervention in order to gain maximum advantage of adopting this approach. For this purpose, agent-based systems are most suited as they typically o er features such as adaptability, pro-activeness, negotiation and learning with a certain degree of autonomy to achieve pre-de ned goals. Thus, the aim of the research was de ned as the integration of information derived from real-time inputs in aiding intelligent decisions with a signi cant degree of autonomy. For this purpose, long term management of nutrition and health needs in modern day relief camps was chosen as a practical use case since it is an often over-looked subject compared to immediate disaster management where a lot of work has been done in developing advanced systems. Moreover, given the current scope of things where over 40 million people have been displaced world over as a result of natural disasters, con-

icts or violence (with more than 1.5 million displaced persons registered in Pakistan alone), the research holds great signi cance. In order to enhance the stability, cost-e ectiveness and reliability, the use of cloud services { in particular infrastructure as a service (IaaS), is highly desirable. Hence, this research endeavor tends to focus on methods for analysis of real-time sensory data, and the development of an automated mechanism for integrating the select pieces of information within a cloudbased system using multi-agents.