www.melanexjou.fora.pl

[loiwtjsf]

Parametric model based intelligent forecasting system and its application


To predict the conditions for success, model, model parameters and results in the form of the right to join the register memory to associative memory system. It should be noted that the modules and the weight of the latter network is essentially an artificial neural network. A large number of artificial neural network known as the node from the information processing unit composed of simple, each node to other nodes adjacent to issue suppression or excitation signal, the entire network of information processing nodes through the interaction of all these done. Neural network input variables and output through variable parameters, the complex relationship between the sample set records in each node associated weight (or threshold),[link widoczny dla zalogowanych], the re-input, neural network by calculating, for association, analogy, memory, in order to achieve the current knowledge of the sample compared with the sample function. (D) weights Network: Neural Network Learning and slander J through the sample train to the right value (or Kan value) in the form of stored samples of knowledge. Parameter model for prediction of different integrated weighting. (E) Inference: Based on the knowledge, according to certain search strategy, heuristic reasoning, prediction model for intelligent selection. The micro level, the inference engine is the set of functions, the premise that it will match the facts and rules, will match the rules on the conclusions of the fact that back as a new knowledge base, knowledge base and then updated all the facts in the premise of matching with the rules, until So far no new facts. The system can forward reasoning, hybrid reasoning, reasoning and the connection confidence level reasoning. Join reasoning is the use of artificial neural networks,[link widoczny dla zalogowanych], and its advantage lies in the fact that given the premise of reasoning results provide the total (of course, depends on the correctness of reasoning given by the results of the adequacy and accuracy of the facts), to avoid the The production rule reasoning the fact that there may be inadequate because the results of the defects can not be provided to ensure continuous intelligent forecasting system. (F) to explain the module; to the reasoning process of inference and reasoning to interpret the results. For example, from which a conclusion to explain reasoning rules, the premise of the rule is verified by what method, the inference engine and the main reason the whole process of reasoning show the path to the user. (2c) Database: Storage of the historical and forecast data and related information. It is linked with the Foxbase2.1. (Viii) Knowledge Development Module: its function is to assist the knowledge acquisition and knowledge base of the structure. The module has to create, modify, insert and other functions, the rules can enter the consistency test for the knowledge base to expand, update, query, modify, provides a possible means. (Ix) Translation dictionary: the Chinese version of intelligent prediction system, all rules of reasoning displayed by the character display section, and in the process of practical reasoning, facts, rules,[link widoczny dla zalogowanych], conclusions are in English, a link for this design English translation in the dictionary, the first letter of the dictionary to retrieve, can be full-screen edit, modify, delete, insert, the user easy to use. (J) interactive interface: as an open,[link widoczny dla zalogowanych], intelligent forecasting system, to facilitate the design and maintenance knowledge, user design a friendly interactive interface, and is equipped with a hot key activated system help file. 4, table and plant equipment downtime intelligent plant equipment downtime predicted a direct impact on production quality and efficiency. How to correctly predict downtime Taiwan, for the combined 40 of Vibration Engineering Volume 6, management organization of production, a reasonable deployment of equipment and staff to ensure the normal order of factory production is critical. Thick solid line in Figure 3 shows a large plant equipment downtime sets monthly time series data when the input of the system. First start the forecasting model selection inference, inference engine, respectively, by forward reasoning and the connection reasoning, are combined in the structural model chosen. Restart structure model, using stepwise regression to extract the exponential trend term, when the confidence level of 95 taken when the power of exponential trend extracted entry for 8774-2f {Trend possible cycle, when the confidence level of 95 taken, the extract of the cycle period of one year Trend 2943.5sin (2 ~ t / lZ); in the rest of the residual series modeling,[link widoczny dla zalogowanych], suitable model for the AR (12). so. downtime when the modular structure of Taiwan model ,173794-8774-2t4-1082 .4 sinEZ ~ r (t a 5) / lZ] 4-E one. _. 26-l + o.162-2 one. _ (a 3 +0-045 324 for a 4 - ... 4-0 * 115 a l24-n, ● I Health j3 factory equipment failure after this model year (12 months) to predict when the station downtime, the results shown in Figure 3, dashed

相关的主题文章：


[link widoczny dla zalogowanych]

[link widoczny dla zalogowanych]

[link widoczny dla zalogowanych]