【摘要】 探讨了可以抵抗回归数据中出现的X与Y空间的多个异常点稳健回归估计方法,介绍了最小一乘法与LTS估计,并利用MATLAB模拟数据,分析以LS,LTS和最小一乘法三种方法对三组模拟数据进行分析比较,表明LTS方法比其他两种的抗粗差效果要好,具有很强的稳健性。更多还原

【Abstract】 The authors explore some robust estimates to resist multiple outliers from both sides and introduce Least absolute deviation and LTS estimator. Also,they compare three simulated data as demonstration. The result shows LTS regression can resist outliers better than the others and it is more robust.更多还原