Informatics Publishing Limited

S. Vigneshwari ¹, M. Aramudhan ²

Affiliations
1 Department of Computer Science and Engineering, Sathyabama University, Chennai, IN
2 Perunthalaivar Kamarajar Institute of Engineering and Technology, Karaikal, Tamilnadu, IN

Abstract

Ontology is the best way for representing the useful information. In this paper, we have planned to develop a model which utilizes multiple ontologies. From those ontologies, based on the mutual information among the concepts the taxonomy is constructed, then the relationship among the concepts is calculated. Thereby the useful information is extracted. There is multiple numbers of ontologies available through the web. But there are various issues to be faced while sharing and reusing the existing ontologies. To resolve the ambiguity which exists, when comparing two concepts are semantically similar, but physically different, an approach is proposed here to index and retrieve the documents from two different ontologies. The ontologies used are WordNet and SWETO ontology. The results are compared based on semantic annotation based on RMS and hashing between the cross ontologies using Rabin Karp fingerprinting algorithm. Also the datasets are trained to yield better results.

Keywords

Concept Similarity, Information Extraction, Hashing, Ontological Relationship, Semantic Annotation, Training the Ontology.

Full Text

   

S. Vigneshwari ¹, M. Aramudhan ²

Affiliations
1 Department of Computer Science and Engineering, Sathyabama University, Chennai, Tamilnadu, IN
2 Department of Information Technology, Perunthalaivar Kamarajar Institute of Science and Technology, Karaikal, Tamilnadu, IN

Abstract

XML data format has become an important one nowadays, because of its compatibility and manageability. XML formats are suitable for both structured data as well as unstructured data. Keyword search can be integrated with XML query processing for querying several XML documents at the same time. Processing XML queries with keyword search is supported by both RDBMS and XML databases. Both the databases are compared as well as web page personalization is also discussed.

Keywords

XML Query Processing, KMP, Information Retrieval, Relational Databases, DST, XML Table.

Full Text

     

Ajith Markose ¹, S. Vigneshwari ¹, S. Gowri ¹, D. Usha Nandhini ¹

Affiliations
1 Faculty of Computing, Sathyabama University, Chennai - 600119, Tamil Nadu, IN

Abstract

Objectives: A privacy-preserving mechanism for public auditing of shared data in cloud storage has been proposed. This boosts up the effectiveness of the verification task which is meant for auditing multiple tasks. It also reduces the response time and auditing time and thereby improves data integrity. Methods/Analysis: A privacy preserving methodology has been proposed which sustains the social interaction and examination on the data which is being mutually shared across the cloud. In scrupulous, ring signatures have been utilized to enhance the verifiability of the computed metadata and to improve the accuracy of the group data analysis. The proposed system maintains the secrecy of the mutual data. The confidentiality of the specific user in the group is ensured by data filtering mechanism. This mechanism masks the user’s private data from being accessed publicly across the cloud. The proposed system also supports multi-group audits simultaneously. Findings: A distinct privacy preserving mechanism is rarely available in the cloud storage especially for shared data. Also the personal information should not be disturbed by public verifiers. The ring mechanism shares only the verified information instead of sharing the entire file. This improves the integrity of the confidential data. The mechanism boosts the potency of substantive multi-group analysis to support the entire data cluster. This improvises the real time cloud data distribution. The identity of the signer is traceable by the group owner. Novelty/Improvement: Only registered users can login to the cloud. This prevents the unauthorized access to the cloud. Data is secured during cloud upload. Other users in the group have no permission to modify the data. Except the signer other users have got read-only permission.

Keywords

Auditing, Authenticators, Batch Auditing, Potency, Privacy, Shared Information.

Full Text

   

S. Kalpana ¹, S. Vigneshwari ²

Affiliations
1 Department of Computer Science and Engineering, Sathyabama University, Chennai - 600119, Tamil Nadu, IN
2 Faculty of computing, Sathyabama University, Chennai - 600119, Tamil Nadu, IN

Abstract

Objective: The main objective is to implement multi view point similarity to perform document comparisons that use the concept of clustering. Methods/Analysis: The main task of data mining is clustering which is used to group or select objects which are similar to one another. Data mining divides whole document into meaningful clusters and analyses data. There are many different types of clustering methods like hierarchical clustering, partitioned clustering and data grouping may be based on distance, viewpoints, Euclidean distance etc, Of these, the current system uses single view point similarity. This type of single view point similarity has some disadvantages. The main disadvantage is it does not use full set of document data so that detailed comparison measures cannot be revealed. In the future system multi viewpoint similarity is used to overcome the above disadvantage. Findings: The multi view point similarity method is used to overcome the disadvantages mentioned under the analysis. This method compares similarity between the multiple documents in detailed manner. The documents have been compared line by line and show the similarity. Then we have enhanced the existing ECSMTP algorithm and it is named as ECSMTP (Enhanced Concept Based Similarity Measure for Text Processing). This algorithm categorizes data from selected documents along with weight age of document, and based on that it forms clusters and calculates the similarity measure. Further in this system different kind of documents were compared like text documents, word, PDF documents etc., but it is not in the existing system. User may select kind of document and comparisons can be made on the selected documents. Clusters were formed and these clusters were compared.

Keywords

Clustering, ECSMTP, Multiviewpoint, Pattern Recognition, Singleview Point

Full Text

   

R. Saranya ¹, S. Gowri ¹, S. Monisha ¹, S. Vigneshwari ¹

Affiliations
1 Department of Information Technology, Sathyabama University, Chennai, IN

Abstract

Objective: The tremendous growth in volume of data is increasing the complexity of data handling. The existing content based search is not only complex and expensive, but also leads to poor processing and unacceptable latency for massive amount of data. Data accessed from large dataset using existing approach will slow down the query response rate. High resource cost is the severe performance bottleneck caused by query operations. The semantic based search can be carried out on the data, In order to reduce the complexity in data handling. The proposed method uses Resource Description Framework (RDF) with Ontology Web Language (OWL). Thus the RDF is used for non domain based search and OWL is used for domain based search. The non domain based search provides the composition of best data services from master RDF. The triplets from different RDF are combined to form master RDF. And the domain based search is one which provides recommendations for the user based on their search query. The main objective of this approach is to provide semantic search using RDF and recommendation using OWL. The search result will be the high similarity result sets as the method uses the combinational RDF to form large master RDF. Statistical Analysis/Methods: The user is provided with two methods of searching which is domain based search and non domain based search. The domain based search is one which provides recommendation for the user based on the search query. And the non domain based search originates the data services from master RDF. Agriculture data set is chosen as a sample data set for domain based search which can be extended further for any other domain. The web services which contain general information about the world and also related web services is chosen as a data set for non domain based search. Findings: The statistical analysis of ontology based semantic search with keyword based search produced the precision value of 0.8 out of 1.0 using the search results obtained from both semantic search and keyword based search. Applications/Improvements: The application can be further improved by adding realistic data sets and increasing the size of the database. The domain based approach is not restricted only to agriculture but it can also be extended to some other applications like Health care, pattern recognition and so on.

Keywords

Correlation, Data Services, Domain, Hazy Semantics, Ontology, Triplets.

Full Text