The Application Of Quotient Space Theory In The Colliery Early Warning

In today's world, coal resource is one of the most important energies. It's also an irreplaceable important role in the economic development. At present, most coal mines in China are high gas concentration and caused the accidents occur frequently. Therefore, prevention, control gas explosion, is the key to achieving safety in coal mines.Coal early warning is the best way to reduce losses in coal mine accident prevention for current security situation.This paper applied the theory of quotient space early warning safety in coal mines. First, on the current profile of coal mine safety production, and coal mine gas-related theory predicted. For coal mine safety production associated with multiple attributes (including wind speed, temperature, carbon monoxide, negative pressure, gas concentration) time series of point features, develop prediction program, the program used in a variety of theoretical knowledge and the solution A detailed analysis of the problem description. Then, of the basic component of granular computing, the basic problem and the main model is summarized and concluded. This article is based on practical problems to be solved for selected commercial space theory. Quotient space theory for treatment of high-dimensional, incomplete, complex, vague, massive data, there are unique advantages. In this paper, the details of commercial space theory, the application of their theoretical knowledge to the time series of pretreatment, the level of granularity from multiple time series analysis of the original deal. How to select the category network model, this paper by comparing the neural network in related models, their advantages and disadvantages, combined with this problems to be solved, developed using constructive neural network (covering algorithm) and the use of different granularity experimental data to train the network. Evaluate this classification by selecting the optimal network model. In the last place, by comparing the regression model, gray model and Markov chain model, the target attributes of the sequence to predict the characteristics of gray Markov chain using multivariate time series properties of the various features of the respective to predict. Finally, the use of these forecasts covering algorithm established in the previous classification model to predict the experiment (prediction of coal mine warning).Aiming at the time series of multiple attributes data related to safety production in the colliery, analysis and process of the time series from multiple levels of granularity by using Quotient Space Theory, to establish the classified network model by Constructive Neural Network (Alternative Covering Algorithm), to train the network by using the experimental data of different granular experimental data. Then, the Gray Markov Model is used to forecast the Multivariate Time Series individually. In the end, to do the prediction experiment using these predicted value on the classified model which was established by Alternative Covering Algorithm(the Colliery Warning Prediction), and the result is satisfying.