Research And Improvement Of An Association Rule Algorithm

Association rules is a very important research direction in Data Mining, which has been widely used in various walks of life. This kind of rule not only can help us to recognize the laws of data within the industry, but also can find and sum up the implication law among mass data, which usually be used as an accurate and reliable data supporting platform for user such as customer and enterprise. In current, database technology is increasingly mature; the scale of data storage has been rapidly expanded, all of that let existing association rules be not adapt to the current data model. So it is very important to seek for more efficient association rules for data mining.The main research contents of the thesis are as follows:The first and foremost, the overview of data mining, this paper briefly introduces the data mining technology generation and development path, emphasis on discussing and analyzing some topics of data mining technology research, such as the current status, hot issue etc.The second, the detailed and comprehensive analysis for data mining, combined with the current situation of scientific research, not only introduced the definition and implementation of the algorithm is given, but also point out its advantages and disadvantages.And then, according to the defects and shortcomings of the algorithm is given, such as low efficiency, simple rules for data mining etc. an improved new algorithm is given, which is a parallel algorithm. The improved algorithm can efficiently cover the shortage of the original algorithm, and show better performance than it.In the last chapter, the parallel algorithm is used in vehicle maintenance4S shop which is becoming widely popular. The parallel algorithm not only seeks for cars fault from the single level, but also does data mining from multi-level car fault in accordance with the association rules. The experimental data finally shows that the parallel algorithm is effective.