Study Of Related Technologies In Rule Mining Based On Historical Data Of Process Object

With the accumulation of data and the intense competition of market, people's need for knowledge and information becomes more and more urgent. Data mining technology can find data hidden in the information and mine the knowledge that has not been discovered. Though people have prior knowledge and experience, they can not determine or predict this information hidden in the data. Therefore data mining provides a lot of help for commercial competition, enterprise production and management, government decision-making, scientific inquiry and other fields. Because of its practical and commercial value, it becomes a hot research and some researchers have come up with many data mining algorithms in recent years.Data mining can not only query and traverse historical data but also identify the potential relationship between these data, thus it can promote the transfer of information. Rule mining comes into being. Rule mining extracts from the database of potential, previously unknown and useful knowledge or rules. As a young and active forefront technology, many emerging engineering fields combines traditional disciplines ,so the rules found and its related technologies has gradually become an important part of intelligent control system. The paper mainly focuses on kinds of related technologies and the comparison of that. Rule mining as an important issue of data mining, it has attracted the widespread concern by the industry in recent years. Because of its great application value and potential theory meaning, all the countries in the world have invested a great deal of manpower, material and financial resources to do a deep research.In this paper, we conduct a study on rule mining and its related technology according to the complexity of industrial production processes, strong correlation, nonlinear, and the characteristics of uncertainty, combined with historical data in process industry, we focus on the rough lattice. Rough lattice is a combination of concept lattice and rough set. Concept lattice and rough set are effective tools in data analysis and knowledge extraction. The Rough Lattice theory that is put forward by this paper to overcome the shortcomings of concept lattice's excessive accuracy and rough set's incomplete and makes a combination of both. So it makes up for their respective deficiencies in data mining and extends the rough set's application fields. In this paper, according to CARCL(Construction Algorithm of Rough Concept Lattice), we combine the two methods'advantages and propose a new method for the construction of Rough Lattice based on compressed matrix (CM_CARCL). The new method can solve the redundancy of construction process.To verify the efficiency of CM_CARCL, the subject does a detailed study of cement production, and applies this algorithm to the entire production line and rule mining. In order to reduce the number of nodes in the rough lattice structure, the program also selects these nodes with eclipse technology to simplify the construction process. The application of compressed matrix greatly reduces space complexity and implementation of the program modules.The experiment can be run well.With the application of CM_CARCL (Compressed Matrix Based Construction Algorithm of Rough Concept Lattice)algorithm in the process industry, we can get the corresponding rules. According to these dug the rules, people can easily identify the relationship between the various process links, then make appropriate regulation and control on the properties of certain objects in the process industry, finally we can achieve expected results.