Wael M. S. Yafooz

Towards automatic column-based data object clustering for multilingual databases

The amount of data in all computer applications is growing tremendously. As a result, the organization of the huge data is crucial. Recently, many researchers consider clustering as one of the important approaches in handling data for wide range of research domains. The examples include Topic Detection and Tracking (TDT), Multilingual Document Clustering, Multilingual News Clustering, Text Clustering and Web Record. Normally, data clustering is time consuming and challenging since they involve heavy programming or scripting. In online news, data clustering analysis is very much needed as the nature of the news across the globe is dynamically changing in every second. The news can come from any web sources in the form of multilingual news. This paper proposes system architecture for an automatic data object clustering in multilingual database for online news, web record and text mining. The architecture provides an overview of a virtual scheme that handles data objects within the database tables as part of the database management system. The proposed technique architecture will provide the platform for quick extraction, data arrangement, data grouping based on pattern similarities. Thus, it will improve query processing performance in multilingual databases without the need to code or script for interface programming. This is the first attempt to apply the data clustering technique prior to data extraction in any database application in the form of semi-structured and structured data (web record).

Challenges and issues on online news management

Recently, the Internet usage spread in all areas of life. Online news is among the popular articles on the Internet, which occupies a large portion of online information. The online news will be viewed almost every second in order to follow the evolution of any desired global events. There are many organizations or political parties employ agents for tracking news by grouping the event. Therefore, news clustering is helpful and worthy for many researchers and online news readers in order to view events from multiple perspectives. Additionally, it can be used in online news summarization, topic detection and tracking for extracting and detecting new events or topics in the news articles. The news extraction can be applied on news articles in the form of monolingual or multilingual. On the other hand, news aggregation is the most important method for scrawling and collecting events based on topics or categorization. This paper investigates the challenges and issues that relate to online news research. The discussions include the overview of system architectures, online news techniques, and a few related computer applications for the above mentioned online news areas.

Managing unstructured data in relational databases

A relational database is a basic repository for many businesses, with its robust data structure for retrieving, organizing, and managing data. However, despite its data structure characteristic, a massive amount of data it contains remains unstructured. These unstructured data affects query processing performance and contributes to the difficulty of the user to manage or retrieve the data. Many attempts have been made to reorganize or directly process these data. In this paper, discusses methods of managing unstructured data in the relational database management system. And show the significance of managing these data. Furthermore, the difference in managing such data between relational and NoSQL databases is highlighted. This study will help developers and researchers in managing unstructured data and in addressing important issues that affect query processing which otherwise meaningless if those were not well managed.

Interactive Big Data Visualization Model Based on Hot Issues (Online News Articles)

Big data is a popular term used to describe a massive volume of data, which is a key component of the current information age. Such data is complex and difficult to understand, and therefore, may be not useful for users in that state. News extraction, aggregation, clustering, news topic detection and tracking, and social network analysis are some of the several attempts that have been made to manage the massive data in social media. Current visualization tools are difficult to adapt to the constant growth of big data, specifically in online news articles. Therefore, this paper proposes Interactive Big Data Visualization Model Based on Hot Issues (IBDVM). IBDVM can be used to visualize hot issues in daily news articles. It is based on textual data clusters in textual databases that improve the performance, accuracy, and quality of big data visualization. This model is useful for online news reader, news agencies, editors, and researchers who involve in textual documents domains.

Model for Automatic Textual Data Clustering in Relational Databases Schema

In the last two decades, unstructured information has become a major challenge in information management. Such challenge is caused by the massive and increasing amount of information resulting from the conversion of almost all daily tasks into digital format. Tools and applications are necessary in organizing unstructured information, which can be found in structured data, such as in relational database management systems (RDBMS). RDBMS has robust and powerful structures for managing, organizing, and retrieving data. However, structured data still contains unstructured information. In this paper, the methods used for managing unstructured data in RDBMS are investigated. In addition, an incremental and dynamic repository for managing unstructured data in relational databases are introduced. The proposed technique organizes unstructured information through linkages among textual data based on semantics. Furthermore, it provides users with a good picture of the unstructured information. The proposed technique can rapidly and easily obtain useful data, and thus, it can be applied in numerous domains, particularly those who deal with textual data, such as news articles.

Shared-Table for Textual Data Clustering in Distributed Relational Databases

High-performance query processing is a significant requirement of database administrators that can be achieved by grouping data into continuous hard disk pages. Such performance can be achieved by using database partitioning techniques. Database partitioning techniques aid in splitting of the physical structure of database tables into small partitions. A distributed database management system is advantageous for many businesses because such a system aids in the achievement of high-performance processing. However, massive amount of data distributed over network nodes affect query processing when retrieving data from different nodes. This study proposes a novel technique based on a shared-table in a relational database under a distributed environment to achieve high-performance query processing by using data mining techniques. A shared-table is used as a guide to show where the data should be saved. Thus, the efficiency of query processing will improve when data is saved at the same location. The proposed method is suitable for news agencies and domains that rely on massive amount of textual data.

Future trends in managing extracted information

Web technology is currently used in all daily activities and is considered a backbone of life. The amount of information continuously increases and grows, specifically that of unstructured information that has no rules or constraints. Such information is difficult to handle and thus requires organization and management before it can be useful. Information extraction techniques are efficient methods of converting unstructured documents into structured data. Attempts have been made to extract structured information that can be used with small amounts of textual data. However, for large amounts of data such as those found in the World Wide Web, the amount of extracted information is huge, and the relationships between extracted information are difficult to determine. Studies that focus on managing extracted information are few. In this paper, we present an overview of the recent studies on managing unstructured information, information extraction and managing extracted information. Managing extracted data using our proposed model for the rapid extraction and clustering of unstructured data for back-end applications in low-level of relational database systems is highlighted. This paper is intended for researchers interested in information extraction management and its applications.