Tag Archives: Clustering.

Application of Cluster Analysis: A Case Study of Customer Buying Behaviour Data (Published)

Despite the wide variety of techniques available for grouping individuals into market segments on the basis of multivariate survey information, clustering remains the most popular and most widely applied method. Clustering is a popular and widely used method for identifying or constructing data based market segments. Over decades of applying cluster analysis procedures for the purpose of searching for homogenous subgroups among consumers, questionable standards of using the techniques have emerged one of such is the black-box approach ignoring crucial parameters of the algorithm applied or the lack of harmonization of methodology chosen and data conditions. This research work is all out to capture: which standard of application of cluster analysis have emerged in the academic marketing literature, compare their standards of applying the methodological knowledge about clustering procedures and delineate sudden changes in clustering habits. These goals are achieved by systematically reviewing some data-driven segmentation studies that apply cluster analysis for partitioning purposes

Keywords: Black-Box Approach, Clustering., Data-Driven Market Segmentation And Homogeneous Subgroup.

Web Data Mining: Views of Criminal Activities (Published)

Web data mining discovers valuable information or knowledge from the web hyperlink structure, page content and usage data. Along with the swift popularity of the Internet, crime information on the web is becoming increasingly flourishing, and the majority of them are in the form of text. A major challenge facing all law-enforcement and intelligence-gathering organizations is accurately and efficiently analyzing the growing volumes of crime data. Detecting, exploring crimes and investigating their relationship with criminals are a big challenge to the present world. The evaluation of the different dimensions of widespread criminal web data causes one of the research challenges to the researchers. Criminal web data always offer convenient and applicable information for law administration and intelligence department. The goal of crime data mining is to understand patterns in criminal behavior in order to predict crime anticipate criminal activity and stop it. This paper describes web data mining which includes structure mining, web content mining, web usage mining and crime data mining. The occurrences of criminal activities based on web data mining process is also presented in this paper. The presented information on different criminal activities can be used to reduce further occurrences of similar incidence and to stop the crime.

Keywords: Classification, Clustering., Crime Control., Crime data, Pattern Analysis, Web Mining

Novel Energy Efficient Election Based Routing Algorithm for Wireless Sensor Network (Published)

Sensor nodes close to the sinks will deplete their limited energy more rapidly than other sensors, since they will have more data to forward during multihop transmission. This will cause network partition, isolated nodes and much shortened network lifetime. Thus, how to balance energy consumption for sensor nodes is an important research issue. In recent years, exploiting sink mobility technology in WSNs has attracted much research attention because it can not only improve energy efficiency, but prolong network lifetime. In this paper, a modified  Election based Protocol, which employs the  decision  of  selecting  cluster  heads  by  the  sink  is  based  on  the  associated additional energy and residual energy at each node. Besides, the cluster head selects the shortest path to reach the sink between the direct approach and the indirect approach  with  the  use  of  the congested link.  Simulation results demonstrate that our algorithm has better performance than traditional routing algorithms, such as LEACH.

Keywords: Clustering., Life time, Multipath Routing, Packet Loss, Wireless Sensor Networks

Web Data Mining: Views of Criminal Activities (Published)

Web data mining discovers valuable information or knowledge from the web hyperlink structure, page content and usage data. Along with the swift popularity of the Internet, crime information on the web is becoming increasingly flourishing, and the majority of them are in the form of text. A major challenge facing all law-enforcement and intelligence-gathering organizations is accurately and efficiently analyzing the growing volumes of crime data. Detecting, exploring crimes and investigating their relationship with criminals are a big challenge to the present world. The evaluation of the different dimensions of widespread criminal web data causes one of the research challenges to the researchers. Criminal web data always offer convenient and applicable information for law administration and intelligence department. The goal of crime data mining is to understand patterns in criminal behavior in order to predict crime anticipate criminal activity and stop it. This paper describes web data mining which includes structure mining, web content mining, web usage mining and crime data mining. The occurrences of criminal activities based on web data mining process is also presented in this paper. The presented information on different criminal activities can be used to reduce further occurrences of similar incidence and to stop the crime.

Keywords: Classification, Clustering., Crime Control., Crime data, Pattern Analysis, Web Mining

EFFECTS OF TECHNOLOGICAL CAPABILITIES, INNOVATIONS AND CLUSTERING ON THE PERFORMANCE OF FIRMS IN THE NIGERIAN FURNITURE INDUSTRY (Published)

The paper evaluates the effects of Technological Capabilities, Innovations and clustering on the performance of firms in furniture making industry in Southwestern Nigeria. The aim is to recommend policy measures to enhance the innovative performance of the furniture makers. The research covered Lagos, Oyo, Ondo and Ekiti States because of the predominant of the industry in these selected locations. The sample population consisted of 319 furniture makers. The research instruments were questionnaire and personal observation approaches. The questionnaire was administered to furniture makers and elicited information on the effects of Technological Capabilities, Innovations and Clustering on the performance of firms in furniture industry in Southwestern Nigeria. Personal observation was used to obtain more information on the industry. Both descriptive and inferential statistical techniques were employed for data analysis. The result shows positive impact of technological capabilities, innovations, and clustering on the performance of the firms on new furniture products produced monthly through adaptation or modification on office furniture, cabinet, upholstery, beds, doors among others. Furniture makers benefit immensely from clustering notably in the area of sharing furniture experience from one another. The ideas of adaptation of furniture were obtained from brain storming of the master furniture makers with their colleagues and apprentices, while catalogue, photograph, magazine and their creativity were instrumental to minor modification of furniture products. Interaction between institutions and Furniture Makers needs to be strengthened to avail them of access to technical support services which institutions can render to the industry. So also financial institutions should finance the industry without interest rate.

Keywords: Clustering., Furniture Industry, Innovations, Technological Capabilities

A HYBRID AND PERSONALIZED ONTOLOGY RANKING MODEL USING U-MEANS CLUSTERING AND HIT COUNT (Published)

Semantic Web is an extension of current Web which offers to add structure to the present Web. Ontologies play an important role in Semantic Web development and retrieval of relevant ontology. Ontology is being represented as a set of concepts and their inter-relationships relevant to some knowledge domain. As the number of Ontology repositories are more on Semantic Web, the problem of retrieving relevant ontologies of the scope arises. Even though there are Semantic Web search engines available, a major problem is that the huge number of results returned and which gives overhead to the searcher to find their need by themselves after going through the long list. This makes time consumption in search and creates dissatisfaction. One solution for this problem is that of maintaining the history of already analyzed, highly relevant and quality results in a log, which can used quickly to respond to the users of the similar type. This places highly relevant results analyzed and stored on the top list when results are presented to the searcher. Personalization and ranking takes care of these approaches. Another solution is the integration of clustering approach which helps in retrieving results from the history or log faster. This paper proposes a hybrid approach that creates the log and retrieves from log when the query is known and there are sufficient entries in the log. This approach imparts convenience to users and reduces the time complexity in finding their relevant needs.

Keywords: Clustering., Ontology, Ontology Ranking, Personalization, Semantic Search, Semantic Web