Data Storage In any industry or field information development seems to have experienced the following process
File ---> Relationship algebra
However, with the in-depth of informationization, the information structure is getting more complex, and the relationship algebra cannot accommodate the management of complex information. So build a knowledge base, it has become a way to have no way.
Build a knowledge base, mainly need to solve the following problems
1 Change and Detection of knowledge:
a) Increase and delete knowledge, need to ensure that the integrity of the original data is not destroyed
b) Query of knowledge, the query of knowledge can be seen as an analysis process, and new knowledge will not be generated during this analysis. If we use a collection of knowledge, aunt is called Knowledge set. When we enter knowledge, set its knowledge set to a. When we exhaust all the queries, set the knowledge set to B, A will never equal B (because the knowledge of the knowledge is an analysis process).
Under normal circumstances B is included in A, B is not included in A. The knowledge loss is a knowledge set C, then c = a-b.
This is actually a reasoning process, in the SQL era, is based on the relationship algebraic reason engine. Through the ratio, it can be found that the storage method, and management method, and the mode of management directly determines the size of the knowledge set C. Only by using a way to allow the partial error in the way, it can make C smaller as much as possible.
2 visualization problem
a) visualization of knowledge content
b) Visualization of reasoning and query processes
Therefore, the construction of the knowledge base management system can be divided into two parts:
1 Select a good data structure and try to make C smaller. In other words, the data structure cannot be made, the knowledge that has not been entered, becomes incompatible with the entered knowledge.
2 Select a variety of reasoning systems to analyze knowledge in the knowledge base
To do it, start with the data structure first
Problem, according to Tasci language, the value (meaning) of the relationship between the elements of the set X is required to be defined in addition to the elements in another collection y. The selection of the data structure of the knowledge base is actually how to choose this collection Y.
This Y is XML Schema in XML, in the relationship algebra, this Y is the dependency master code. Because the definition of dependencies in SQL is simple, the SQL itself is smaller, so the success of SQL
The problem with SQL is that although it defines dependencies, it is not possible to further define dependencies, and this work is handed over to the application.
The problem now is that for complex data structures, how to choose a good collection Y.
