Research On Data Mining In The Process Of Coal Mine Safety Training Game

In the production of coal mine, mine disaster occurred frequently and caused se-rious personal safety and property loss, and the nonstandard operations of human be-havior are the main cause, the coal mine safety training can improve the quality ofperson training safely and effectively, so it is an important means to prevent coal mineaccidents. Through the interaction of the user with the realistic simulation of coalmine safety production simulation environment, and also with the people’s perceptionand cognition for object which they contact, coal mine safety training game is usefulfor inspiring the user’s thinking, comprehensively accessing to a variety of spatialinformation, logical information and professional knowledge which contained in theproduction environment, and also improving the effect of training. So how to improvethe user’s training effect of the training game is the most concerned problem to man-agers. While users in the game will produce a large amount of data, and these data canbe used to analyze the users behavior, so this thesis uses the data mining technologyin the coal mine safety training game to analyze the users operation behavior, betterimprove user’s training and game management.In order to find the association of user behavior in the coal mine safety traininggame, and to provide basis to management decision-making, and better to service us-ers, using association rule to mining the relationship between users behavior. Throughthe analysis of advantages and disadvantages of mining association rules algorithm,this thesis raises an correlation analysis method based on improved genetic algorithm,,At first using improved genetic algorithm to find the frequent itemsets, and then re-generating strong association rules. According to the coal mine safety training gamedata have the feature of diversity and complexity, this thesis uses the real numbercoding mechanism, designs the fitness function based on association rules, with em-phasis on improving the genetic operation. It improves the anction of crossover andmutation. Comparing with the traditional genetic algorithm, the improved algorithmhas a fast convergence speed, high efficiency. And comparing with the classical Ap-riori algorithm of association rules, the algorithm’s running time is short.After the analysis of coal mine safety training game data, these data are classi-fied according to the different of acquisition and storage mode, the obtained data arepreprocessed, and these value data are sorted out to build the model of user behaviordata, then coding for the establishment of data model, converting the data to map data which can be used for the association rule mining based on genetic algorithm, usingthe map data model to data mining association.Finally this thesis designs the Interface of management system of coal minesafety training game using the soft of Qt, it can displays these data tables and datamining results, and then analyses the strong association rules, and puts forward rea-sonable scheme and provides valuable basis for management decisions.