人才流动的滑动平均马尔科夫链预测模型离散模拟  

Moving Average Talent Flow Simulation Based on Markov Chain Prediction Model

作　　者：郑峰[1] 

机构地区：[1]滁州学院经济与管理学院,安徽滁州239000

出　　处：《统计与决策》2017年第3期32-35,共4页Statistics & Decision

基　　金：国家自然科学基金青年项目(71502002);安徽省人文社科重点研究基地项目(SK2016A0678);安徽省哲学社会科学规划青年项目(AHSKQ2016D108);安徽省高校优秀青年人才支持计划项目

摘　　要：文章以支持向量机SVM模型进行修正,结合滑动平均处理获得相应的人才流动预测样本序列的均值和方差,基于SVM基础上的滑动平均马尔科夫权重归一化处理。结果显示,经过平均滤波器迭代后的样本序列参与滑动平均处理,有利于更高精度意义的加权马尔科夫链预测,且通过滑动平均处理,来获得适合加权马尔科夫链分析的样本预测过程,对于长时序和大样本研究具有更好的适应性。This paper modifies the support vector machine SVM model, combines with the moving average value and variance of the flow of talent obtain corresponding prediction samples, based on the SVM model of the sliding average power normalization processing that based on Markov. The results show that the sample sequence average filter after iteration in moving average, is conducive to higher precision of the weighted Markov chain prediction, and by moving average processing, to obtain suitable for the analysis of weighted Markov chain prediction process is better for a long time sequence and large sample study adaptability.

关 键 词：加权滑动处理 马尔科夫链 滑动平均处理 平均滤波器 

分 类 号：O21[理学—概率论与数理统计]