In a file management environment, each application has a specific data file related to it, containing all the data records needed by the application. Records stored in a sequential file structure may be accessed sequentially, or they may be accessed directly via an index (and then sequentially) using an indexed sequential access method. Records stored in a data file structure may be accessed directly without using an index.
The traditional data file organization led to many problems, including data redundancy, data inconsistency, data isolation, data integrity, security, and application/data dependence. Storing data in data files that are tightly to their applications resulted in organizations having hundreds of applications and data files, with little or no coordination among the applications and files, and no overall plan for managing corporate data.
A database, which is a logical group of related files, eliminates the problems associated with a traditional file environment. In a database data are integrated and related so that one set of software program provides access to all the data. Therefore, data redundancy, data inconsistency and data isolation are minimized and data can be shared among all users of the data. In addition data integrity and security are increased, and application and data are independent of one another.
The database approach does have disadvantages. Databases are expensive and require time and effort to program. Also, databases do provide security for corporate data, but once inside a database, a hacker can cause tremendous damage.
The hierarchical model rigidly structures data into an inverted 'tree' in which records contain a key field and a number of other fields. All records have only one "parent", and each parent may have many "children". Therefore, the hierarchical structure is characterized by one-to-many relationships among data. In the network model, records can be linked to more than one parent, allowing many-to-many relationship among the data. The relational model uses tables to capitalize on characteristics of rows and columns of data that are consistent with real world business situations.
The main advantage of the hierarchical and network database models is processing efficiency. The hierarchical and network structures are relatively easy for users to understand because they reflect the pattern of many (but not all) real-world business relationships. In addition, the hierarchical structure allows for data integrity to be easily maintained.
Hierarchical and network structures have several disadvantages. These designs have low flexibility and are programming intensive, time consuming, difficult to install and difficult to remedy if design errors occur. Nor do they support ad-hoc, English-language-like inquiries for information.
The advantages of relational databases include high flexibility in regard to ad hoc queries, power to combine information from different sources, simplicity of design and maintenance, and the ability to add new data and records without disturbing existing applications. The disadvantages of relational databases include their relatively low processing efficiency.
In multidimensional databases, data are stored in arrays. Similar to tables in the relational database model, arrays group related information in columns and rows. However, multidimensional databases typically consist of at least three dimensions. Due to problems depicting more than three dimensions, most examples artificially limit the dimensions to only three, depicting the resulting database as a cube. Dimensions are the edges of the cube, and represent the primary view of the business data.
Data warehousing approaches can range from simple, the data mart, to complex, the enterprise data warehouse. These approaches differ in scale and complexity.
A data mart is a scaled down version of a data warehouse that focuses on a particular subject area. The data mart is usually designed to support the unique business requirements of a specific department or business process. Because a data mart takes less time to build, costs less, and is less complex than an enterprise data warehouse, it is appropriate when a company needs to improve data access in a targeted area, such as the marketing department.
The enterprise data warehouse provides an enterprisewide, consistent and comprehensive view of the company, with business users employing common terminology and data standards through out the firm. The warehouse reconciles the various departmental perspectives into a single, integrated corporate perspective.
Data mining extracts previously unknown, predictive information from data warehouses. Data mining tools are sophisticated, automated algorithms to discover hidden patterns, correlations, and relationships among organizational data. These tools are used to predict future trends and behaviors, allowing businesses to make proactive, knowledge-driven decisions.
Text mining applies data mining to nonstructured or less structured text files. Text mining helps organization find the "hidden" content of documents across previously unnoticed divisions, and group documents by common themes.
No comments:
Post a Comment