期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2016
卷号：94
期号：1
出版社：Journal of Theoretical and Applied
摘要：Organizations mainly rely on data and the mechanism of dealing with that data on cloud computing. Data in an organization has multi security levels, which is classified depending on nature of the data, and the impact of data on the organization. The security procedures which used for protecting data usually be complicated, and it had a direct and indirect influence on the usability level. This study aims to establish a model which has an ability to classify data dynamically according to the security form low till high levels. The security level classified it into five levels based on the policies and classification method. The purpose of classification is to apply a complex security procedure on data which has a high security level larger than data which has a low security level. It also has a potential to segregation an illegal data from the legal to support usability in system. Finally, several experiments have been conducted to evaluate the proposed approaches. Several experiments have been performed to empirically evaluate two feature selection methods (Chi-square (χ2), information gain (IG)) and five classification methods (decision tree classifier, Support Vector Machine (SVM), Na�ve Bayes (NB), and K-Nearest Neighbor (KNN) and meta-classifier combination) for Legal Documents Filtering The results show that all classifiers perform better with the information gain feature selection methods than their results with Chi-Square feature selection method. Results also show that Support Vector Machine (SVM) outperforms achieve the best results among all individual classifiers. However, the proposed meta-classifiers method achieves the best results among all classification approaches.
关键词：Information System Security; Classification Of Data; Big Data; Neural Language Processing
