Research On Data Mining Algorithms For Intelligence Analysis From Some AWACS Radar Detection

AWACS mission when recording mass data contains a wealth of valuable information,in order to enhance the combat capability of information,we need a large number of AWACS intelligence data analysis.How to extract information from massive amounts of data in order to analyze intelligence service,how to improve the efficiency and accuracy of the information and intelligence massive data mining,to improve the quality of intelligence and enhance the combat capability is very important,has been the traditional data analysis facing major problem.Emerging cloud computing technology for the above problem provides a new and effective way to solve,using distributed computing techniques to achieve large-scale parallel computing,greatly enhance the data processing capabilities,dynamic and ready to quickly provide a large number of cheap computing power,thus,cloud computing technology for AWACS massive data mining,is expected to break through the bottleneck of traditional data mining,mining intelligence effectively enhance efficiency.According to a certain type of early-warning radar intelligence analysis on the massive demand intelligence data mining,Hadoop cloud computing framework from the start,relying on its distributed framework,the framework research operation mechanism Map/Reduce computing,the data mining algorithms based on Map/Reduce is proposed,and then tThe algorithm is applied to detect early-warning intelligence information analysis platform.The major work of the thesis is as following aspects:Firstly,in order to provide a better quality of intelligence mining data,data normalization in data preparation phase is studied in the thesis,and the two major problems of the traditional normalization method is analyzed.And in accordance with some problems,Hive tool normalization methods of fast Map/Reduce operation combined with Hadoop platform is proposed,the normalization method is fast and highly scalable fully demonstrated by the test of the AWACS single radar data.Secondly,this thesis focuses on the study of mining algorithm,found the reason that restricting the speed of the algorithm by analyzing the process of the classic Apriori algorithm of association rules,Map/Reduce-based data mining algorithm is proposed.The algorithm transformes Apriori algorithm into the Map/Reduce model to achieve Apriori parallel transformation,and it uses the way to compress the original transaction set to achieve a highly scalable of the suitable Map/Reduce algorithm for cloud computing environment.Finally,Set up early-warning detection intelligence information analysis platform for improved data mining algorithms for testing,verify the reliability of the algorithm,accuracy and efficiency,and then by an AWACS radar intelligence real fly data mining analysis,the detection of this type of early warning aircraft leaks appear strong association rules affect larger factors.