Study On Material Cost Forecast Of Coal Mine Based On Data Mining

Material consumption has a wind variety and huge number in coal miningprocess of coal production enterprises. Material cost is generally about20%of thetotal cost,35%of direct cost. Meanwhile, material input of coal mine is also the mostcontrollable project, through to accurately predict the material cost, effectively controlof material consumption, coal mine may reduce coal production cost and improve thecompetitiveness.Material cost prediction should be based on the scientific mathematical models,to reduce the influence of subjective factors on the quota and other indicators ofdecomposition. Coupled with management information systems, database technologyhas been accumulated a large amount of material in the continuous application of coalmining enterprises, the database consumes historical data. Through data miningtechnology into the material cost coal forecast maximize the value of first-hand dataon the long-term accumulation of material cost data mining, data was found hiddenbehind knowledge, decision-making basis for the material cost management toimprove coal enterprises material cost management. Around this theme, this papercarries out the following research.First, this paper summarizes mines material cost management, data miningtechnology, coal research in data mining, data mining literature in terms of costmanagement and other aspects of the research. Concepts, tasks, processes andmethods of data mining are introduced and explained the advantages of data miningtechnology compared to traditional statistical methods.Then, the composition, characteristics, management status, predicted problems incoal material costs were analyzed. On this basis, a number of factors affecting the coalmaterial costs summarized, combined with material cost projections principles forfactors to quantify the material to build model based on structural equation modelinganalysis of the factors affecting the cost model for coal derived material hereinpredictable cost factors, including production footage, occurrence geologicalconditions, tons of tunneling rate, degree of mechanization, equipment, technical level,the input and output variables to build predictive models of coal material costs toprovide evidence.Then, the paper analyzes the support vector machine learning rules and thetraining process, introduces the particle swarm (PSO) algorithm to optimize the parameters of support vector machine to build support vector regression (PSO-SVR)mine material cost prediction model based on PSO. Coal mine material cost inaccordance with predictions based on data mining process model, preprocess theactual production data in a mine, and perform data mining and transformation ofknowledge. In the forecasting process, I predicted the total cost of material andaccounting for a large proportion of the large material cost and supporting materialcost, and fitted the predicted results with the actual value, tested the relative error. Theresult shows that forecasting effect is satisfied, and I give the new predictive value.Finally, based on the coal related material cost prediction, the author put forwardsome policy recommendations concerning the coal mine enterprise materialmanagement, in requirements planning and purchasing plan, material cost control andmaterial management information, etc.