"Top down" vs. "Bottom up", "direct" vs. "Reverse", "Manage" vs. "Purpose driven" - three pairs of definitions for terms such as "chain analysis, parsing, syntactic parsing," "logical analysis" and "search." In principle, all these terms reflect a similar attitude, and the difference between them is simply that they are drawn from different subareas of computer science and artificial intelligence (parsing, a system with a set of rules of search engines and systems, aimed at solving problems and etc.)
These contrapositions can be illustrated by the paradigm of search. The main task of any search is to identify the route by which you will move to this position to your goal. If you start the search with the current position and will continue it until natknetes the desired result - so-called direct search or search for a bottom-up. If you mentally put yourself in the place where you want to find yourself on a search and determine a route, heading in the opposite direction, ie where you really are at the moment - is the search in the opposite direction, or by looking down. Please note that by defining the route as a result of a search, you still will get to their goal. Despite the fact that now you're moving forward, this is not a direct search, as search has been carried out previously, in the opposite direction.
The same contrast could be considered an example of systems with embedded rules. Imagine that the rule consists of a set of background and a set of consequences. When the system determines that all the background certain rules are satisfied, this rule is called and executed (executed there, each caused by a rule depends on the particular system). After that, a knowledge base put approval obtained as a result of the rules and comply with the relevant transaction. This process occurs as described above, irrespective of whether the system of direct or inverse logical analysis. To illustrate the differences between them, should separately consider the procedure for activating the rule. Caused only activated rules. Direct logical analysis (upward) when the system adds the new data, they are compared with all the background of all the rules. If the data are consistent with the antecedent of rule, this rule is activated (if it is not yet activated), and if you picked all the background specific rule, it is called. Claims resulting from the implementation of the rules entered in the knowledge base and are regarded as new data are compared with the background and can cause activation of the call and additional rules. In a reverse logical analysis (top down) when you add these rules are not activated. When the system receives a request, it compares with all the consequences of all rules. If the request coincides with the result, this rule is activated, and all of his background are considered as secondary queries, and can cause the activation of additional rules.
When a query matches are not limited to approval of a knowledge base, it receives a response and if the request comes from a background, it is believed that he satisfies the latter. When all the background of a rule are met, the rule is called and executed. When the rules of being a response to requests that it activated, and the other background are satisfied and may be caused by corresponding rules. Please note that the call and execution of rules is always happening in the direct sequence, but unlike the direct analysis of a chain is when a rule is activated.
These contrapositions can be illustrated by the paradigm of search. The main task of any search is to identify the route by which you will move to this position to your goal. If you start the search with the current position and will continue it until natknetes the desired result - so-called direct search or search for a bottom-up. If you mentally put yourself in the place where you want to find yourself on a search and determine a route, heading in the opposite direction, ie where you really are at the moment - is the search in the opposite direction, or by looking down. Please note that by defining the route as a result of a search, you still will get to their goal. Despite the fact that now you're moving forward, this is not a direct search, as search has been carried out previously, in the opposite direction.
The same contrast could be considered an example of systems with embedded rules. Imagine that the rule consists of a set of background and a set of consequences. When the system determines that all the background certain rules are satisfied, this rule is called and executed (executed there, each caused by a rule depends on the particular system). After that, a knowledge base put approval obtained as a result of the rules and comply with the relevant transaction. This process occurs as described above, irrespective of whether the system of direct or inverse logical analysis. To illustrate the differences between them, should separately consider the procedure for activating the rule. Caused only activated rules. Direct logical analysis (upward) when the system adds the new data, they are compared with all the background of all the rules. If the data are consistent with the antecedent of rule, this rule is activated (if it is not yet activated), and if you picked all the background specific rule, it is called. Claims resulting from the implementation of the rules entered in the knowledge base and are regarded as new data are compared with the background and can cause activation of the call and additional rules. In a reverse logical analysis (top down) when you add these rules are not activated. When the system receives a request, it compares with all the consequences of all rules. If the request coincides with the result, this rule is activated, and all of his background are considered as secondary queries, and can cause the activation of additional rules.
When a query matches are not limited to approval of a knowledge base, it receives a response and if the request comes from a background, it is believed that he satisfies the latter. When all the background of a rule are met, the rule is called and executed. When the rules of being a response to requests that it activated, and the other background are satisfied and may be caused by corresponding rules. Please note that the call and execution of rules is always happening in the direct sequence, but unlike the direct analysis of a chain is when a rule is activated.
Комментариев нет:
Отправить комментарий