Mydhili K. Nair

Prediction of Heart Disease Using Classification Based Data Mining Techniques

Data Mining is an interesting field of research whose major objective is to find interesting and useful patterns from huge data sets. These patterns can be further used to make important decisions based on the result of the analysis. Healthcare industry today generates huge amount of data on a day to day basis. This data has to be analysed and hidden and meaningful patterns can be discovered. Data mining plays a promising and significant role in this aspect. Data Mining techniques can be used for disease prediction. In this research, the classification based data mining techniques are applied to healthcare data. This research focuses on the prediction of heart disease using three classification techniques namely Decision Trees, Naive Bayes and K Nearest Neighbour.

‘CloudCop’: Putting network-admin on cloud nine towards Cloud Computing for Network Monitoring

Computer Network Monitoring is an evergreen field of challenges. The recent advent of Cloud Computing, in the realm of ‘New Generation Networks’, opens a whole new field of envisaging and implementing at least two of its aspects, namely ‘Software as a Service(SaaS)’and ‘Platform as a Service(PaaS)’ in the field of Network Management(NM). In this paper, we present the design and implementation of a Network Monitoring Framework oriented towards Cloud Computing which we call ‘CloudCop’. The NM framework uses Web Services (WS) and Service Oriented Computing (SOC) to implement SaaS, while PaaS is provided by providing some of the features such as WS marshalling, integration, storage and persistence. Its data storage schema is made very generic to be applicable for other applications thus being in tune with the vision of SOC. It also employs a whole gamut of technologies such as WS, SOC, Mobile Agents(MA), Simple Network Management Protocol(SNMP) and is made scalable and robust enough to absorb new pieces such as Rule Engines and Ontology based Service Discovery which are planned in the next implementation phase. Therefore, this paper attempts to provide directions on the applicability of Cloud Computing for NM with a case study and implementation results paving the way for the users of such applications, network system administrators (network-admin)truly elated and therefore on ‘cloud nine’!

Net Mobile-Cop: A hybrid intelli-agent framework to manage networks

More than a decade ago, MA(Mobile Agents) were introduced for NM(Network Management) of distributed computer networks in tandem with the design paradigm of Code- Shipping the agents to remote node, where they are programmed to collect, analyze and process data locally. However, even today, most networks employ SNMP (Simple Network Management Protocol), an inherently Client-Server based stable and proven protocol, which uses the paradigm Data-Shipping. It suffers from the major drawback that when the network traffic increases, the manager is overloaded due to excessive processing. Real synergy could be achieved if we adopt a hybrid approach that brings the best of both paradigms. In this paper we present Net Mobile-CopHybrid NM Framework, prototyped on Aglet Mobile Agent System and AdventNet SNMP Package, which imbibes this synergy. We also introduce a novel method to dynamically configure managed nodes using intelligent agents. We also provide quantitative evaluation leading to useful tips to the sys-admin as to when to toggle between SNMP and MA usage.

Temperature monitoring using Sun SPOTS applied to vermiculture

In our ongoing research work, we focus on how to amalgamate technology with organic farming. In this paper, we present a case study of using temperature sensor-actuators to monitor the temperature of the habitat of the worms used for vermiculture, a very vital stream in the realm of organic farming. ‘Vermiculture’ meaning artificial rearing or cultivation of worms is a scientific process used to make vermicompost, an excellent, nutrient-rich fertilizer and soil conditioner. Vermicompost is a crucial ingredient in organic farming. The most common red worms used in vermiculture are very sensitive to temperature making it imperative to monitor the temperature of the habitat of these worms. We have done a field study of the worm habitats at University of Agricultural Sciences (U.A.S), Gandhi Krishi Vigyan Kendra (G.K.V.K), Bangalore. The temperature was monitored using Sun SPOT sensor-actuators donated by SUN Microsystems for academic research. In this paper, we present our experimental results.

Effect of Outlier Detection on Clustering Accuracy and Computation Time of CHB K-Means Algorithm

Data clustering is one of the major areas of research in data mining. Of late, high dimensionality dataset is becoming popular because of the generation of huge volumes of data. Among the traditional partitional clustering algorithms, the bisecting K-Means is one of the most widely used for high dimensional dataset. But the performance degrades as the dimensionality increases. Also, the task of selection of cluster for further bisection is a challenging one. To overcome these drawbacks, we incorporate two constraints namely, stability-based measure and Mean Square Error (MSE) on the novel partitional clustering method, CHB-K-Means algorithm. In the experimental analysis, the performance is analyzed with respect to computation time and clustering accuracy as the number of outliers detected varies. We infer that an average clustering accuracy of 75 % has been achieved and the computation time taken for cluster formation also decreases as more number of outliers is detected.

Development of Fractional Genetic PSO Algorithm for Multi Objective Data Clustering

Clustering is the task of finding natural partitioning within a data set such that data items within the same group are more similar than those within different groups. The performance of the traditional K-Means and Bisecting K-Means algorithm degrades as the dimensionality of the data increases. In order to find better clustering results, it is important to enhance the traditional algorithms by incorporating various constraints. Hence it is planned to develop a Multi-Objective Optimization MOO technique by including different objectives, like MSE, Stability measure, DB index, XB-index and sym-index. These five objectives will be used as fitness function for the proposed Fractional Genetic PSO algorithm FGPSO which is the hybrid optimization algorithm to do the clustering process. The performance of the proposed multi objective FGPSO algorithm will be evaluated based on clustering accuracy. Finally, the applicability of the proposed algorithm will be checked for some benchmark data sets available in the UCI machine learning repository.

Anveshana — Search for the right service

In an era witnessing exponential growth in number of web services in similar domains being published in internet, redundancy is an inevitable issue. In this context, Quality of Service which is offered by these services forms the differentiating factor. Several works have been published describing different approaches to discover best in class services and it continues to be a hot research area. In this process, service providers who provide ordinary quality of service lose out to competition and eventually dissolve. Hence customers will have to choose service providers who provide better quality than they actually need and end up paying more money. In order to address this issue, we have proposed a new approach named Anveshana aiming to create a Win-Win result for both customer and service provider. Anveshana considers quality and priority requirements of customer and provides a relevance score to each applicable service and helps in rationally selecting the right service in an unbiased manner. This forms basis of further research on customer driven service selection leveraging on quality of service based computations.

Enhancement of K-Means algorithm using ACO as an optimization technique on high dimensional data

Clustering is a distribution of data into groups of similar objects such that the objects in a group will be similar (or related) to one another and different from (or unrelated to) the objects in other groups. The concept of clustering applications is particularly in the context of information retrieval and in organizing web resources. The objective of clustering is to find out information and in the present day context, to locate most relevant resources. In data mining, K-Means clustering is a method of cluster analysis which aims to partition n observations into k clusters in which each observation belongs to the cluster with the nearest mean. Though the K-Means is one of the best clustering algorithms, the quality is based on the starting condition and it may converge to local minima. There is not much of work done by the researchers to improve the cluster quality after grouping. We have proposed a novel method to improve the cluster quality on high dimensional data set by ant based refinement algorithm. The Ant Colony Optimization algorithm (ACO) is one of the most widely used probabilistic technique for solving computational problems which can be reduced to finding good paths through graphs. An ant is a simple computational agent in the ant colony optimization algorithm. It develops an iterative solution for any problem at hand. The intermediate solutions can be used to arrive at the final solution. The proposed algorithm is tested using data from different domain and the results show that refined initial starting points and post processing refinement of clusters based on ACO can lead to improved solutions in terms of entropy, time taken and accuracy of clusters.

Agent Based Web Services with RuleML for Network Management

This Event-Condition-Action(ECA) rule description language for Web Services(named WS-ECA) is the latest buzzword in the realm of autonomous, intelligent Agent Based Web Service Oriented Computing (AWSOC). This rule description language supports interactions among service oriented, heterogeneous devices in a pervasive computing environment. It is based on ECA rules, which are embedded in distributed service devices and are triggered by events from internal or external factors. In this paper, we report our ongoing efforts to use DamlRuleML to define the ECA rules for a hybrid SNMP and Mobile Agent (MA) Based Network Management System (NMS) we have built in-house. Our work is a novel approach embedding ECA Rules into AWS (Agent Based Web Services) which is used to enable pervasive network management.

Digital Privacy and Data Security in Cloud-Based Services: A Call for Action

The emergence of cloud-based services over the past two decades has created a paradigm shift in computing. Fueled by the universal reach of internet services, all the knowledge areas have openly embraced digitization of data storage, computation and communication. Social connectivity has made a village out of our digital world and we are more connected than ever before. With this multitude of facilities, comes the challenge of dealing with security and privacy concerns of our socially and digitally connected world. This work discusses the real-world scenario and highlights these concerns from the perspective of each stakeholder for our holistic understanding. The need for trustable solutions for secure cloud-based solutions is highlighted. Possible solutions that have been proposed include user identity protection by data masking and secure data sharing, incorporation of tools and practices for building a secure development life cycle, and developing holistic, universal privacy laws that are technically correct and auditable in real time. We call upon fellow researchers to take up research in these open research areas to defend the fortress of digital privacy and data security of cloud-based web applications.