基于模糊信息粒化与马尔科夫修正的瓦斯浓度预测

Prediction of Gas Concentration Based on Fuzzy Information Granulation and Markov Correction

基于模糊信息粒化对原始数据进行处理,利用支持向量机(SVM)以及差分进化算法(DE)建立预测模型,根据马尔科夫链进行残差修正,预测瓦斯浓度变化趋势。通过实验表明,该方法相较于直接对粒化数据进行SVM预测有了较大的效果提升,此外,对原始数据进行了FIG处理,实现了降维,算法运行时间得到减少,算法效率得到提升。

The original data is processed based on fuzzy information granulation, and the support vector machine （SVM） and differential evolution algorithm （DE） are used to establish the prediction model, and the residual correction is carried out according to the Markov chain. Predict the trend of gas concentration change. Experiments show that this method has improved compared with the direct of the granulation original data,in addition, FIG treatment, achieve dimensionality reduction algorithm running time is reduced, the algorithm efficiency is improved.