Informatics Publishing Limited

G. Bhavani ¹, S. Sangeetha ¹, S. Sivakumari ¹

Affiliations
1 Department of Computer Science and Engineering, Avinashilingam Institute for Home Science and Higher Education for Women, Coimbatore, IN

Abstract

A huge amount of data stored on the Internet will be useful and helpful only if it is accessed as information, not as pure data. Nowadays Big data overcomes several issues such as searching, analysing, sharing, storage, transfer, visualization and querying. Among these issues, semantic retrieval is a huge issue. In order to avoid these problems, Hadoop Distributed File System (HDFS) is proposed. HDFS performs semantic analysis over the volume of documents (Big data) to find the best matched source document from the collected set of source documents for the same virtual document. In the hadoop file system, the semantic analysis is done using Dual Walk based Ranking model for providing best matched documents and the resulting documents are filtered by making use of Top K algorithm based on the frequency of the entities in the source document. But, the existing system still has issues with the ontological indexing concept and hence the accuracy of semantic information retrieval is reduced. In order to overcome this ontological indexing concept is focused to retrieve highly relevant and semantic information. Ontology based information retrieval increases the most relevant information by filtering the unrelated terms in the documents. The documents are clustered based on the Ontology and the input query is examined for semantics and expanded using domain Ontology. Thus the accuracy of the semantic information is increased and searching complexity is reduced significantly. From the experimental result, the conclusion decides that the proposed system is better than the existing system.

Keywords

Big Data, HDFS, Information Retrieval, Ontology.

Full Text

     

M. Aasha ¹, S. Sivakumari ¹

Affiliations
1 Dept. of Computer Science & Engineering, Avinashilingam Institute for Home Science and Higher Education for Women, Coimbatore, IN

Abstract

Gait is the manner in which a person walks and has gained much importance in the recent past in surveillance systems. This method uses the concept of extracting the features from the video sequence. which are used to identify the individual. In this work, along with the most effective parts and more informative less effective parts, which are extracted due to the effect of various cofactors, shape features are also considered for recognition. This work utilizes only the most informative movable parts with fixed movement as the shape of the parts tends to change with motion. The performance of this method is compared against gait recognition using PSO without utilizing the shape features. The experimental results shows that the proposed method of PSO utilizing shape features shows better performance metrics when compared to the recognition using PSO without considering the shape features.

Keywords

Gait Recognition, Multi-Objective PSO, Shape Feature.

Full Text

     

S. Sivaranjani ¹, S. Sivakumari ¹

Affiliations
1 Department of Computer Science and Engineering, Avinashilingam Institute for Home Science and Higher Education for Women, Coimbatore - 641108, Tamil Nadu, IN

Abstract

Objective: The main objective of this research is to reduce the burden of crime investigators by identifying the series of crimes happening at different places. And also, this work aims to reduce the investigation time by grouping similar crimes happened in different places based on its behavior with the consideration of the class imbalance problem. Methods: In this research, Majority Weighted Class Oversampling (MWCS) method is introduced which overcomes the problem of class imbalance problem. It is introduced to handle class imbalance problem by identifying the hard to learn information which is named as minority class samples from the major class samples. And also in this work, the Incremental Clustering (IC) is introduced which can handle the insertion and deletion operations where the existing methodology called graph cut clustering algorithm cant handle these problems. The proposed methodologies deal with the class imbalance problems effectively and also the modification processes over the partitioned graphs are supported well than the existing researches. Results: The methods used in this work namely MWCS and IC are used to detect the series of crimes by identifying the similarity relationship exists among the crimes happened in different places. The experimental tests conducted were proves that the proposed methodology can leads to well detection of serial residential crimes than the existing methodologies. The experimental results of this work prove that the proposed methodology is improved in terms of all performance metrics called jaccard index, mantel index and the journey distance time. Conclusion: The findings demonstrate that serial residential crimes are identified by clustering them effectively using the methodologies called MWCs and IC and it has high possibility of detection of crimes than the existing methodology.

Keywords

Cut Clustering, Incremental Clustering, Majority Weighted Class Oversampling, Serial Crimes

Full Text

   

P. Amudha ¹, S. Karthik ², S. Sivakumari ³

Affiliations
1 Department of CSE, Faculty of Engineering, Avinashilingam Institute for Home Science and Higher Education for Women, Mettupalayam Road, Coimbatore - 641 108, Tamilnadu, IN
2 Department of CSE, SNS College of Technology, SNS Kalvi Nagar, Sathy Main Road, NH-209, Vazhiyampalayam, Saravanampatti Post, Coimbatore - 641035, Tamil Nadu, IN
3 Department of CSE, Avinashilingam Institute for Home Science and Higher Education for Women, Mettupalayam Road, Coimbatore - 641108, Tamil Nadu,, IN

Abstract

Background: Recently network security is achieved using intrusion detection, in which data mining techniquesare used as a new methodology. The vital features considered is one of the major aspects that affect the efficiency of the Intrusion Detection System (IDS). Methods: The key idea of this work is to propose a feature selection method to discover useful features and to classify user behaviour patterns of system features from the network traffic data using classification approaches. In the process of selecting significant features, the dimensions of data is reduced and the features are sorted by finding the accuracy of each attribute and then selects the best vital features among them based on its accuracy value. Also this work aims to choose a hybrid classifier model (ABC-SVM) based on Artificial Bee Colony (ABC) and Support Vector Machine (SVM) algorithms to construct a perfect IDS using KDDCup'99 dataset. Results: The result analysis indicates that the features selected improve the accuracy rate of ABC-SVM than using all features. Also the hybrid algorithm is better than other traditional algorithms with respect to the performance measures such as detection rate, specificity and training time.

Keywords

Artificial Bee Colony, Classification, Data Mining, Intrusion Detection, Network Security, Support Vector Machine

Full Text

   

R. Praveena Priyadarsini ¹, M. L. Valarmathi ², S. Sivakumari ¹

Affiliations
1 Department of Computer Science and Engineering, Avinashilingam Deemed University for Women, Tamil Nadu, IN
2 Department of Computer Science and Engineering, Government College of Technology, Tamil Nadu, IN

Abstract

Privacy-preservation is a step in data mining that tries to safeguard sensitive information from unsanctioned disclosure and hence protecting individual data records and their privacy. There are various privacy preservation techniques like k-anonymity, l-diversity and t-closeness and data perturbation. In this paper k-anonymity privacy protection technique is applied to high dimensional datasets like adult and census. since, both the data sets are high dimensional, feature subset selection method like Gain Ratio is applied and the attributes of the datasets are ranked and low ranking attributes are filtered to form new reduced data subsets. K-anonymization privacy preservation technique is then applied on reduced datasets. The accuracy of the privacy preserved reduced datasets and the original datasets are compared for their accuracy on the two functionalities of data mining namely classification and clustering using naïve Bayesian and k-means algorithm respectively. Experimental results show that classification and clustering accuracy are comparatively the same for reduced k-anonym zed datasets and the original data sets.

Keywords

Privacy Preservation, Data Mining, K-Anonymity, Feature Subset Selection, Gain Ratio.

Full Text