K. G. Srinivasagan

Bio: K. G. Srinivasagan is an academic researcher from National Engineering College. The author has contributed to research in topics: Computer science & Cloud computing. The author has an hindex of 7, co-authored 14 publications receiving 274 citations.

A New Text Mining Approach Based on HMM-SVM for Web News Classification

Abstract: Since the emergence of WWW, it is essential to handle a very large amount of electronic data of which the majority is in the form of text. This scenario can be effectively handled by various Data Mining techniques. This paper proposes an intelligent system for online news classification based on Hidden Markov Model (HMM) and Support Vector Machine (SVM). An intelligent system is designed to extract the keywords from the online news paper content and classify it according to the pre defined categories. Three different stages are designed to classify the content of online newspapers such as (1) Text pre-processing (2) HMM based Feature Extraction and (3) Classification using SVM. Data have been collected for experimentation from The Hindu, The New Indian Express, Times of India, Business Line, and The Economic Times. The experimental results are based on the news categories such as sports, finance and politics and their accuracies in percentage are 92.45, 96.34 and 90.76 respectively. These results are very good compared to that of other text classification methods.

An improved dynamic data replica selection and placement in cloud

Abstract: Cloud computing platform is getting more and more attentions as a new trend of data management. Data replication has been widely used to speed up data access in cloud. Replica selection and placement are the major issues in replication. In this paper we propose an approach for dynamic data replication in cloud. A replica management system allows users to create, and manage replicas and update the replicas if the original datas are modified. The proposed work concentrates on designing an algorithm for suitable optimal replica selection and placement to increase availability of data in the cloud. The method consists of two main phases file application and replication operation. The first phase contains the replica location and creation by using catalog and index. In second phase is used to find whether there is enough space in the destination to store the requested file or not. Replication aims to increase availability of resources, minimum access cost, shared bandwidth consumption and delay time by replicating data. The proposed systems developed under the Eucalyptus cloud environment. The results of proposed replica selection algorithm achieve better accessibility compared with other methods.

Minimizing EndtoEnd Delay using Multipath Routing in Wireless Sensor Networks

Abstract: Some sensitive applications such as volcanic monitoring, fire detection data should be transmitted within a specified delay to the base station. Multipath-GT (Multipath Generalized Topology) model uses an on-demand approach to estimate a delay based on processing time, packet loss rate between two neighbouring nodes. In existing work, if a node or link failure occurs multipath routing didn’t spread traffic over alternate paths. This paper take a view that, when certain nodes and links become over-utilized and cause congestion, proposed work can spread traffic over alternate paths to balance the load over those paths and increase the degree of fault tolerance. The simulation results show that reduces the probability of communication disruption and data loss during link failures.

Cuckoo search algorithm and wind driven optimization based study of satellite image segmentation for multilevel thresholding using Kapur's entropy

Abstract: The objective of image segmentation is to extract meaningful objects. A meaningful segmentation selects the proper threshold values to optimize a criterion using entropy. The conventional multilevel thresholding methods are efficient for bi-level thresholding. However, they are computationally expensive when extended to multilevel thresholding since they exhaustively search the optimal thresholds to optimize the objective functions. To overcome this problem, two successful swarm-intelligence-based global optimization algorithms, cuckoo search (CS) algorithm and wind driven optimization (WDO) for multilevel thresholding using Kapur's entropy has been employed. For this purpose, best solution as fitness function is achieved through CS and WDO algorithm using Kapur's entropy for optimal multilevel thresholding. A new approach of CS and WDO algorithm is used for selection of optimal threshold value. This algorithm is used to obtain the best solution or best fitness value from the initial random threshold values, and to evaluate the quality of a solution, correlation function is used. Experimental results have been examined on standard set of satellite images using various numbers of thresholds. The results based on Kapur's entropy reveal that CS, ELR-CS and WDO method can be accurately and efficiently used in multilevel thresholding problem.

Surveying wearable human assistive technology for life and safety critical applications: standards, challenges and opportunities

Abstract: In this survey a new application paradigm life and safety for critical operations and missions using wearable Wireless Body Area Networks (WBANs) technology is introduced. This paradigm has a vast scope of applications, including disaster management, worker safety in harsh environments such as roadside and building workers, mobile health monitoring, ambient assisted living and many more. It is often the case that during the critical operations and the target conditions, the existing infrastructure is either absent, damaged or overcrowded. In this context, it is envisioned that WBANs will enable the quick deployment of ad-hoc/on-the-fly communication networks to help save many lives and ensuring people's safety. However, to understand the applications more deeply and their specific characteristics and requirements, this survey presents a comprehensive study on the applications scenarios, their context and specific requirements. It explores details of the key enabling standards, existing state-of-the-art research studies, and projects to understand their limitations before realizing aforementioned applications. Application-specific challenges and issues are discussed comprehensively from various perspectives and future research and development directions are highlighted as an inspiration for new innovative solutions. To conclude, this survey opens up a good opportunity for companies and research centers to investigate old but still new problems, in the realm of wearable technologies, which are increasingly evolving and getting more and more attention recently.

Tsallis entropy based multilevel thresholding for colored satellite image segmentation using evolutionary algorithms

Abstract: Cuckoo search based multi-level thresholding is presented by maximizing the Tsallis entropy.Different optimization algorithms are exploited with Tsallis entropy method.Cuckoo based Tsallis entropy was found to be more accurate for colored satellite image segmentation.The feasibility of the proposed approach has been tested on 10 different colored satellite images. In this paper, a new technique for color image segmentation using CS algorithm supported by Tsallis entropy for multilevel thresholding has been proposed toward the effective colored segmentation of satellite images. The nonextensive entropy is a new expansion in statistical mechanics, and it is a recent formalism in which a real quantity q was introduced as parameter for physical systems that presents the long range interactions, long time memories and fractal-type structures. The feasibility of the proposed cuckoo search and Tsallis entropy based approach was tested on 10 different satellite images and benchmarked with differential evolution, wind driven optimization, particle swarm optimization and artificial bee colony algorithm for solving the multilevel colored image thresholding problems. Experiments have been conducted on a variety of satellite images. Several measurements are used to evaluate the performance of proposed method which clearly illustrates the effectiveness and robustness of the proposed algorithm. The experimental results qualitative and quantitative both demonstrate that the proposed method selects the threshold values effectively and properly.