Earticle

In this paper we have presented a loss-less compression technique namely H-HIBASE (compression enhancement of HIBASE technique using Huffman Coding). Due to disk based compression, H-HIBASE supports very large database with acceptable storage volume. Insertion, deletion and update mechanisms on the architecture have been presented and analyzed. The architecture executes query directly on compressed data and it is capable of executing all types of SQL queries. The experimental evaluation has been performed with synthetic and real data. The experimental result has been compared with DHIBASE and widely used Oracle database. We have evaluated the storage performance in comparison with DHIBASE and Oracle database. The storage performance that has been achieved in H-HIBASE is 25 to 40 percent better than the Oracle database for real and synthetic data. The query performance that has been achieved in H-HIBASE is 10 to 25 percent better than that of DHIBASE.

Cluster and outlier detection has always been one of data mining research interests. Numerous approaches have been designed to find clusters and detect outliers in various types of data sets. In this paper, we present our research on analyzing data sets with constant changes. We design approaches to keep track of status of clusters, the movement of data points, and the updated group of outliers. Different from the traditional approaches which are focused on two-dimensional or low-dimensional data spaces, we aim to analyze data sets in multi-dimensional data spaces. We also propose to adjust the clusters and outliers simultaneously, since they are two concepts that are closely related.

Class imbalance is a problem that is very much critical in many real-world application domains of machine learning. When examples of one class in a training data set vastly outnumber examples of the other class(es), traditional data mining algorithms tend to create suboptimal classification models. Researchers have rigorously studied several techniques to alleviate the problem of class imbalance, including resampling algorithms, and feature selection approaches to this problem. In this paper, we present a new hybrid feature selection algorithm dubbed as Class Imbalance Learning using Intelligent Under Sampling (CILIUS), for learning from skewed training data. This algorithm provides a simpler and faster alternative by using C4.5 as base algorithm. We conduct experiments using four UCI data sets from various application domains using five learning algorithms for comparison and five evaluation metrics. Experimental results show that our method has higher Area under the ROC Curve, F-measure, Precision, TP rate and low TN rate values than many existing class imbalance learning methods.

A high security level DBMS requires a formal specification and verification on the security model and top level specification design. The specification and verification towards SQL operations are important especially. In this paper, based on the security model and top level specification, we propose a novel approach to solve the specification and verification issues towards SQL operations. Firstly, we formally define the SQL operations in FTLS; then, we give the definitions of the simple SQL operations and propose a method to verify those simple SQL operations; finally, we transform the verification of the SQL operations in FTLS to the verification of the component simple SQL operations. The process of verification shows that our approach makes a comprehensive specification of SQL operations and simplifies the verification procedure.

Peer-to-peer (P2P) databases are becoming prevalent on the Internet for distribution and sharing of documents, applications, and other digital media. The problem of answering large-scale ad hoc analysis queries, for example, aggregation queries, on these databases poses unique challenges. Exact solutions can be time consuming and difficult to implement, for the distributed and dynamic nature of P2P databases. In this paper, we have presented novel sampling-based techniques for approximate answering of ad hoc aggregation queries in such databases. Computing a high-quality random sample of the database efficiently in the P2P environment is complicated due to several factors: the data is distributed (usually in uneven quantities) across many peers, within each peer the data is often highly correlated and moreover, even collecting a random sample of the peers is difficult to accomplish. To formulate these problems, developed software uses approach, based on random walks of the P2P graph, as well as block-level sampling techniques. We have presented here extensive experimental evaluations to demonstrate the feasibility of our solution. The modules that are included in this project are: 1. Peer-to-Peer Node Construction 2. Random Selection of Node 3. Selection of Records 4. Performance Evaluation

In this paper is developed a model to cluster customers who need to production change continuously. Mere mathematical models to cluster customers lead to ignoring the environment factors in model. Therefore, for Adjusting model to reality, we should use environment factors in model building. The proposed model is flexible against changes of environment and it causes to resulted model can use for clustering the customers of decay able productions. Proposed model perform on customer's information of a textile firm.

In this paper, the new intuitionistic fuzzy rough relational model that expands the traditional model with the ability to represent indiscernible and imprecise information is presented. We utilize the notion of indiscernibility from rough set theory coupled with the idea of membership and non-membership values from intuitionistic fuzzy set theory to represent uncertain information in a manner that maintains the degree of uncertainty of information for each tuple of the original database and also those resulting from queries. First, we design our database using some type of semantic model and then use a variation of the entity-relationship diagram i.e. intuitionistic fuzzy rough E-R diagram. We also define the operators for the intuitionistic fuzzy rough relational database and demonstrate the expressive power of the model through its intuitionistic fuzzy rough relational algebra.