Research On RoughSet And CBR In Weather Forecast

In weather forecast domain, the people proposed the urgent need to the disaster weather forecast. Historical weather data and the forecast experience are often important to the weather forecast result .A multimedia information comprehensive knowledge storehouse,which established with the experience and the historical weather model analysis and so on, will become the primary assistant to make the real-time weather forecast. Therefore to the synthesis weather case base establishment, the retrieval and so on the thorough research has the important theory significance and the project practical value.This article adopts the research technique which the simulation experiment and the theory unifies, excavates the knowledge from the meteorological data, the union experience builds the case base And uses the CBR and rough set technology to the similar retrieval, thus enhances the weather forecast the level.In the article proposed the weather forecast CBR model, has studied the disaster weather case base establishment method, withdraws the case base use by the data mining way compared with the manul establishment case base the good and bad points, proposed the data model and extracts the case the algorithm; Has analyzed the rough set in the weather forecast function, has studied the case base attribute separate method, and through the discernibility matrix method to reduct attribute, reduced the model retrieval time and the spatial expenses.In the article proposed the similar weather model fast retrieval mentality, first uses the roughset theory to remove the redundancy the decision-making table attribute, then uses the BP nerve network and the similar principle realization similar model retrieval. The experimental result indicated that, uses the roughest and the CBR technology to the similar weather retrieval. the speed is quick, the success ratio is high.