Research on Rough set applied in the geological measure data prediction model

Mine geological factors involved in measuring systems are very complex, large amount of data and attribute more. In actual measurement, due to the precision of measuring instruments and measurement operations personnel and other reasons, the data is inevitably flawed, and then has the subsequent impact of the design, production and management. With intelligent technology and development of computer science, mine has an increasingly high demand of geologic measure data and there are more and more methods to deal with the data. In this paper, rough set theory is applied by analyzing the characteristics of geological measure data and the structure of the database, the corresponding model is established, the uncertainty is found from the database of knowledge and abnormal data, and geological measure dataand the noise in the process of knowledge discovery interference is eliminated. As rough set method is easy to execute in parallel, without any prior knowledge of data and automatically select the appropriate set of attributes, can greatly improve the efficiency of knowledge discovery, get rid of excess property, reduce error rates, then has more advantages in processing the mass of the geological measure data and mining a more realistic data than fuzzy sets and neural network method. In addition, it is easier to be proven and tested in the resulting decision rules and reasoning processes than the latter neural network method are, and the results obtained is more easily evaluated and interpreted. Thus, using rough sets to mine the geologic measure data and find the knowledge hidden in the data, and then make the forecast analysis and decision-support for mine production and management, which is more practical.