Research And Design Of Topic Discovery Based On OODA

To relieve the serious situation "information explosion, Data rich, little knowledge", we put forward a kind of topic discovery about OODA decision recycling. It can gain valuable information from large amounts of data rapidly and efficiently. Besides, we can know something about the data set content. We usually divide it from broad and narrow sense. The generalized topic discovery aims to various usual kinds of data set (text, images, audio, video, etc.). The narrow sense topic discovery aims at text data. What we usually say is the narrow sense one. This paper introduces OODA similar to human thinking patterns and regards it as a topic discovery framework, it applies intelligent data mining, text mining, data fusion and knowledge discovery as well as other related thoughts to study and analyze the topic discovery from different angles .People can gain necessary information from a large amount of complicated information quickly and analyze it. Its related methods, models and ideas can be applied to information retrieval, text classification and clustering, public opinion monitor, literature retrieval active information delivery as well as other areas.At present, the topic discovery lacks theme the advanced framework .The pattern, recognition, features extraction and similarity calculation all need the improved algorithm. Adaptation is inadequate, besides, it lacks the topic discovery in special area. To show the users the visual discovery progress and results, we need improve this method to make users to under, evaluate, communicate.This paper not only studies OODA as the topic discovery technical framework also improves the weight of TF * IDF algorithm and removing unused words; What is more, it uses association rules mining to solve the recognition problem compounds; In order to improve field adaptation and accuracy key words extraction, it sets the data set with association rules and the improved algorithm combining TF * IDF thesaurus .And it makes a similarity calculation and studies words ,phrases, text with the improved K - means theme clustering algorithm; Then it uses the visual NetBeans development tools to identity related technologies. Besides, it is convenient to understand and analyze, interact with users by displaying the progress. Furthermore, it can be improved easily. We can use accumulated BBS data of a certain university basis on OODA framework topic discovery and find out hot issues in the BBS data.All in all, after lots of experiments and being analyzed, this topic discovery is effective. It can analyze the topic discovery questions effectively and divide them abstractly. It is easy to understand, so it is suitable to the analysis and designation of the topic discovery .Some innovative technical analysis method gain some better results and are improved than before.