Data warehouse systems include data warehouses (
Data warehouse
), Data mining (
Data Mining
) And data center library (
Data repository
). The data warehouse, data mining and data center library will be described in detail below.
database
The data warehouse is the topic, integrated, different time, stable data set, used to support the decision development process in business management. That is, the data warehouse is a process, the process organizes and stores data from the perspective of history, and can integrate data analysis. Briefly, the data warehouse is a large database that stores all of the company's business data, for example, online transaction processing (
OLTP
) The integrated data acquired in the system may reside in many different data sources. These data sources may be documentation, hierarchical database, network structured database, reverse list database, relational database (for example
SQL Server
) Or more common hybrid systems consisting of the above systems. Data warehouse can assist in decision support and online analysis process (
OLAP
)application.
Data mining
Data mining is found from large databases or data warehouses and extracts information or knowledge hidden in it. The aim is to help analyst finding the association between data, discovering the ignored elements, and this information is very useful for predicting trends and decision behavior. The general process of data mining is as shown
1
Indicated.
Figure
1
General process of data mining
The general process of data mining includes five steps. (
1
Pre-processing data: Collect and purify information from the data source, and store it, usually store it in the data warehouse. (
2
Model search: Using data mining tools to find models in data, this search process can be automatically executed by the system, searching for original facts upwards to find some link between them, and can also join the user interaction process, active by analyst Send it, find the correctness of the assumption from top to bottom. Many tools may be used for a search process for a problem. For example, neural networks, rule-based systems, based on instance-based reasoning, machine learning, statistical methods, etc. (
3
Evaluation output results. Generally speaking, the search process of data mining needs to repeat multiple times, because when the analyst evaluates the output results, they may form some new issues or require a more fine query for a certain aspect. (
4
) Generate the final result report. (
5
Interpretation results report. The results are explained, and the corresponding business measures are taken according to this result. This is an artificial process.
Data center library
In order to provide a more successful data warehouse and data transaction function, the integration of data yuan is the current most important task. When the primary data conversion service specifies, and online analysis process (
OLAP
) After extending the open information model of the memory, the open design inspection phase begins. These important extensions refer to increasing several information models to the data center library, allowing developers to provide more optional compatible products and data warehouse systems.
The data center library provides a generic location that can be used to store the relationship between objects and objects. The object-oriented information is described by using some software tools. The architecture of the data center library is as shown
2
Indicated.
Figure
2
Data center library architecture
The reason for creating a data warehouse is because the company is more dependent on the collection information from the information system. Therefore, a information data warehouse is required for the company's operation. Customers also want to access the company's data after licenses. Data warehouse is the information provided by business analysts, which is hard to get in the past business database. In most cases, the company will transfer historical data from the business database to the backup system, which makes users unable to analyze data, difficult to make competitive decisions. To better manage data, maintain data consistency, and demand for data warehouses from enterprises to analyze data. The data warehouse allows you to share data between the individual departments of the company, providing more accurate and complete information for business faster, better business decisions. The implementation of the data warehouse is ultimately completed by many support tools, including
OLAP
Service, data conversion service,
Pivottable
Service, English query services, etc. The steps in general, design, and creating data warehouses are: Determine user needs, design, and create databases, extract, and load data warehouses.
