A tetrahedral data model for unstructured data management

Abstract

This paper proposes a tetrahedral data model for unstructured data management. The model defines the four components of unstructured data including: basic attributes, semantic characteristics, low-level features and raw data on its four facets, and the relations between these components. The internal implementation structure of the model and the data query language are designed and briefly introduced. This model provides a unified, integrated and associated description for different kinds of unstructured data, and supports intelligent data services such as associated retrieval and data mining. An example is given to demonstrate how to use the model for describing and manipulating data from a sample video base.With the rapid development of information technology, the quantity of unstructured data has increased dramatically, and accounts for about 80% of the total data in the world. Unlike structured data that is stored in traditional database systems and is described by normal data structures, unstructured data lacks the explicit semantic structure necessary for computerized interpretation [1]. The major types of unstructured data include text, graphics, image, audio and video data. Computer software requires annotations from human beings or machines in order to properly classify and manipulate this data. Hence, the management of unstructured data is far more complex than structured data, and traditional relational database systems are limited to the management of structured data.The management of unstructured data is a new focus in the data management area. Search engines, such as Google and Yahoo, provide fast searching based on keywords; content-based information retrieval technology implements queries using lowlevel feature matching; content management systems provide deep search functions for certain kinds of unstructured data; and ontology-based data query has been proposed. Moreover, traditional database management systems, such as Oracle and DB2, have been extended to support the storage and manipulation of unstructured data. However, most of the above systems are only adequate for handling limited kinds of unstructured data, and only provide limited data retrieval methods. To manage and manipulate large amounts of unstructured data, associated retrieval that is based on multiple retrieval methods and other intelligent data services such as data analysis and data mining are required. Since data in these systems is described either by descriptive text or by lowlevel features, the above intelligent data services cannot effectively be supported. Therefore, in order to

Development and Regulation of Medical Internet

“…These applications have improved the old health care services in varying degrees. [13] Due to the characteristics of medical services in China determines the current domestic Internet equal medical applications to medical resources and improve the efficiency of the service as a starting point, therefore, make an appointment, medicine, electricity and other auxiliary medical application in improving efficiency, convenient side effect is obvious. [14] Wearable devices, cloud storage and service, and big data management and to the health of the residents of the combination of rehabilitation provides a new means and support, so residents can be more easily for health management and rehabilitation.…”

AUDR: An Advanced Unstructured Data Repository

“…Based on above analysis of unstructured data, the uniform data model named Tetrahedral Data Model was recently proposed by Wei Li and Bo Lang [6]. The tetrahedral data model presented in this paper characterizes unstructured data based on these four aspects.…”

“…First a uniform data model named Tetrahedral model [6] for unstructured data is proposed to unify the description of various types of unstructured data and intrmsIc link between the components. Based on the Tetrahedral model, we propose a scalable architecture, which not only adapts to the emerging unstructured data, but also supports the distributed massive unstructured data storage and computing under multiple operating system platforms [7].…”

“…Data model is a formal representation based on data abstract, and query language is the formal representation based on data operation abstract. The underlying data model of AUDR named Tetrahedral Data Model was proposed in [6] by Wei Li and Bo Lang. Unstructured data can be described by a tetrahedral, with facets standing for basic attribute, semantic feature, low-level feature and raw data respectively.…”

Analysis and Establishment of Big data Model

“…(2) Research on the processing flow of keyword search requests and semantic retrieval request, and the processing of strategy and implementation for analysis and visualization of the request for Big data services is made through in-depth analysis of Big data retrieval, analysis and visualization requirements. Practical projects are made to verify the case [2] . At the same time, in order to optimize the ability of Big data services, study of the search results ranking optimization algorithm and persistent cache mechanism based on mixed prefetch are carried out respectively from the two aspects that improve the accuracy of search results and service efficiency.…”