P4 - Product Model | Computer Science (lecture)
Productive model of knowledge representation is one of the most common. Knowledge representation wardrobe using rules-productions is in some ways similar to the inference rules of logic models. This allows using the efficient output productions and, in addition, due to the natural process of reasoning similar human data models clearly reflect knowledge.
In systems products knowledge represented by a set of rules of the form: "if A then B". Where A and B can be understood as a "situation - action", "cause - effect", "condition - output", etc. Often a rule-production record using a logical mark the passage:
It should not be confused rule-production ratio and logical passage. The fact that the interpretation depends on the product that are left and right of the mark logical passage. Often when A is understood some information structure (eg, a frame) and under in - some action that is its transformation wardrobe (transformation). The logical interpretation of the considered expression imposes wardrobe rigid restrictions on the content of the symbols A and B, are to be regarded as PPF. Therefore, the notion of logical product wider passage. As an example, consider a rule taken from the knowledge base of the expert system MYCIN [53] intended to diagnose infectious diseases.
Body of the rule consists of four facts connected union "and". If all four facts take place, the result will be true to the rules. Each rule is associated a number that takes values between -1 and 1, which expresses wardrobe the degree of probability of outcome and called confidence coefficient. This means that the system is inaccurate and unreliable evidence. Some methods work with imprecise knowledge discussed in Section 4.2.
3)
If during activation of the first four rules of evidence will prove true, the knowledge base is put a new fact which seems triplet object-attribute-value. One advantage of preserving evidence wardrobe in the form of triplets is to improve the efficiency of searching in the knowledge base, so that it is possible to organize the facts in objects and attributes.
where Rnj - j-ID products in n-set of products, PR-priority rules for production, wardrobe Sun - a prerequisite for the application of kernel production, which is a predicate, which is activated when performing core products, Ace - pislyaumova product that defines the actions and procedures to be perform after the kernel output.
The set of production rules form the basis of rules, each of which represents the detached fragments of knowledge about the solvable problem. wardrobe Psychologists call these fragments chankamy (from Eng. Chunk). It is believed that chan - a unit of objectively existing knowledge vyhlyadilyayetsya man in cognition of the world.
The premise of the rule is often seen as a model. This is some information structure that defines a generalized situation (condition, condition, etc.) of reality in which the rule is activated. Working memory (global database) reflects the specific situation (state conditions) occurring in the environment. Information structure that represents the specific situation of the environment in the working memory, called the image.
The interpreter implements the logical conclusion. The process output is cyclical and is called the search for a model. Consider it in a simplified form. Current status of domain modeled is reflected in the working memory as a set of images, each of which is represented by means of facts. Memory is initialized facts describing the problem. Then selected those rules for which samples submitted preconditions of rules compared with the images in working memory. Data rules constitute a conflict set. All rules are in conflict set may be active. According to Favorites conflict resolution mechanism activates one of the rules. Action contained in the output of the rules leads to changes in the state of working memory. Further output control cycle is repeated. This process ends when there will be rules, conditions are compared with images of working memory (Figure 3.2).
Explain the operation of the production process of a simple example sorting wardrobe a string wardrobe consisting wardrobe of the letters a, bi s [77]. Terms sort presented in Figure 3.3. If the pattern specified wardrobe prerequisite rules comparable to the sorted part of the line, the rule is activated. As a result of this subexpression, which coincided with the condition rules subexpression is replaced with the final part of the rules.
Considering the source and sorted rows as the initial and final state of the task, and production rules as operators that transform one state to another, we reach the conclusion that the search of solutions
Productive model of knowledge representation is one of the most common. Knowledge representation wardrobe using rules-productions is in some ways similar to the inference rules of logic models. This allows using the efficient output productions and, in addition, due to the natural process of reasoning similar human data models clearly reflect knowledge.
In systems products knowledge represented by a set of rules of the form: "if A then B". Where A and B can be understood as a "situation - action", "cause - effect", "condition - output", etc. Often a rule-production record using a logical mark the passage:
It should not be confused rule-production ratio and logical passage. The fact that the interpretation depends on the product that are left and right of the mark logical passage. Often when A is understood some information structure (eg, a frame) and under in - some action that is its transformation wardrobe (transformation). The logical interpretation of the considered expression imposes wardrobe rigid restrictions on the content of the symbols A and B, are to be regarded as PPF. Therefore, the notion of logical product wider passage. As an example, consider a rule taken from the knowledge base of the expert system MYCIN [53] intended to diagnose infectious diseases.
Body of the rule consists of four facts connected union "and". If all four facts take place, the result will be true to the rules. Each rule is associated a number that takes values between -1 and 1, which expresses wardrobe the degree of probability of outcome and called confidence coefficient. This means that the system is inaccurate and unreliable evidence. Some methods work with imprecise knowledge discussed in Section 4.2.
3)
If during activation of the first four rules of evidence will prove true, the knowledge base is put a new fact which seems triplet object-attribute-value. One advantage of preserving evidence wardrobe in the form of triplets is to improve the efficiency of searching in the knowledge base, so that it is possible to organize the facts in objects and attributes.
where Rnj - j-ID products in n-set of products, PR-priority rules for production, wardrobe Sun - a prerequisite for the application of kernel production, which is a predicate, which is activated when performing core products, Ace - pislyaumova product that defines the actions and procedures to be perform after the kernel output.
The set of production rules form the basis of rules, each of which represents the detached fragments of knowledge about the solvable problem. wardrobe Psychologists call these fragments chankamy (from Eng. Chunk). It is believed that chan - a unit of objectively existing knowledge vyhlyadilyayetsya man in cognition of the world.
The premise of the rule is often seen as a model. This is some information structure that defines a generalized situation (condition, condition, etc.) of reality in which the rule is activated. Working memory (global database) reflects the specific situation (state conditions) occurring in the environment. Information structure that represents the specific situation of the environment in the working memory, called the image.
The interpreter implements the logical conclusion. The process output is cyclical and is called the search for a model. Consider it in a simplified form. Current status of domain modeled is reflected in the working memory as a set of images, each of which is represented by means of facts. Memory is initialized facts describing the problem. Then selected those rules for which samples submitted preconditions of rules compared with the images in working memory. Data rules constitute a conflict set. All rules are in conflict set may be active. According to Favorites conflict resolution mechanism activates one of the rules. Action contained in the output of the rules leads to changes in the state of working memory. Further output control cycle is repeated. This process ends when there will be rules, conditions are compared with images of working memory (Figure 3.2).
Explain the operation of the production process of a simple example sorting wardrobe a string wardrobe consisting wardrobe of the letters a, bi s [77]. Terms sort presented in Figure 3.3. If the pattern specified wardrobe prerequisite rules comparable to the sorted part of the line, the rule is activated. As a result of this subexpression, which coincided with the condition rules subexpression is replaced with the final part of the rules.
Considering the source and sorted rows as the initial and final state of the task, and production rules as operators that transform one state to another, we reach the conclusion that the search of solutions
No comments:
Post a Comment