| With the Bayesian networks gradual used in a number of fault diagnosis, the advantages of the Bayesian network as the basis theory in the probabilistic reasoning embodied in the increasingly evident, for the research and application of the Bayesian network learning is also widely developed, but for the Bayesian network modeling and probabilistic reasoning and some other common computing operations due to large amount of complex structures such factors as lower research and application of the Bayesian networks the efficiency of staff, reducing the Bayesian network computing accuracy of the data, from a certain point of affecting the development of the Bayesian network.The Bayesian network development platform is intended to improve the efficiency of network research and application staff and the efficiency of the Bayesian network computing accuracy, so put forward. The platform uses eclipse as a development platform, java as a development language, using the RCP + GEF technology development and uses the Bayesian network's relevant knowledge as a theoretical basis. In order to make the platform more intelligent, the platform also integrates a number of algorithms, such as hill-climbing algorithm, K2 algorithm, makes only the data link between the view data or loss in the Bayesian Network graphical recovery case operation more efficient, convenient, and convenient the Bayesian network application development.In the research phase, proposed a new method of data recovery, can be automatically repair the missing data of the Bayesian network, proven accuracy rate of about 80%, and has been integrated into the platform, the application of the Bayesian network development platform can improve the Bayesian network staff efficiency. |
