Prediction Of Mining Surface Settlement Based On Ant Colony Optimization Algorithm For Support Vector Machine

Mining surface settlement problem is a slow and long process.With the settlement of the accumulation,the fact that the total amount exceeds the maximum threshold will be serious landslide and other disasters.In order to prevent mining surface settlement from the disasters,it is very impotent that monitors settlement data and control their settlement continuously.Currently there is much prediction method for mining,but there is larger error,slow convergence problem with traditional forecasting methods.Therefore,there is great significance to explore new and effective method for predicting the settlement of mine surface.Research on mining surface settlement prediction based on support vector machine at home and abroad,I will apply ant colony optimization to improve the performance of support vector machine prediction model.First,I described the basic theory,mathematical models and their characteristics about ant colony optimization and support vector machine;Secondly,I apply the gridding of ant colony optimization to select parameters of support vector machine;I adopt C # to build predictive models in Visual Studio;Finally,I use Fibonacci weighting method to process date of mining surface settlement,then to build mining surface settlement prediction model based on ant colony optimization algorithm for support vector machine and to comparative analysis with mining surface settlement prediction model based on traditional support vector machine.The result show that the error is varies 0.0015 m from 0.0028 m on 12 th,varies0.0015 m from 0.0025 m on 22 th,varies 0.0018 m from 0.0028 m on 24 th in our way,the error is varies 0.0055 m from 0.0086 m on 12 th,varies 0.0060 m from 0.0078 m on22th,varies 0.0058 m from 0.0081 m on 24 th in traditional way.The results show that,the precision in our way is better than in traditional way and predictive value curve closer measured value in our way than in traditional way.Thus,we can get the fact that mining surface settlement prediction model based on ant colony optimization algorithm for support vector machine is feasible and usefulness in the promotion of practical engineering applications.