摘要
对系统抽样的估计方差 ^V1(-ysy) = ( 1M- 1N) 1M- 1 ∑Mj=1( Yij- -Yi)2 ^V2(-ysy) = ( 1M- 1N) 12( M- 1) ∑Mj=1( Yij- Yi,j+ 1)2我们有下列结果:(i) 如果ρk ≠0 ,V(-ysy) = V(-y) ,则E[ ^V2(-ysy)] ≠E[ ^V1(-ysy)] = V(-y) ;(ii) 如果ρk < 0 ,V(-ysy) < V(-y) ,则E[ ^V2(-ysy)] > E[ ^V1(-ysy)]> V(-y) ;(iii) 如果ρk > 0 ,V(-ysy) > V(-y) ,则E[ ^V2(-ysy)] < E[^V1(-ysy)] < V(-y) .其中V(-ysy) = 1K∑ki= 1-Yi - -Y 2 ,V(-y) = ( 1M- 1N) 1N- 1 ∑Ki= 1 ∑Mj=1Yij- -Y 2ρk =∑Ki= 1 ∑M-1j=1( Yij - -Yi)( Yi,j+ 1 - -Yi) + 12 ∑Ki=1( Yi1 - -Yi)2 + ∑Ki=1( YiM- -Yi)2∑Ki=1 ∑Mj= 1Yij - -Yi
Forthe estimated systematic sampling variance ^V 1(-ysy) = ( 1 M- 1 N) 1 M- 1 ∑M j= 1( Yij - 珡Y i)2 ^V 2(-ysy) = ( 1 M- 1 N) 1 2( M- 1) ∑M j= 1( Yij- Yi,j+ 1)2 We obtain thefollowing main theorem (i) if ρk ≠0 ,V(-ysy) = V(-y) ,then E[ ^V 2(-ysy)] ≠E[ ^V 1(-ysy)] = V(-y) ; (ii)if ρk < 0 ,V(-ysy)< V(-y) ,then E[ ^V 2(-ysy)] > E[ ^V 1(-ysy)] > V(-y) ;(iii) if ρk > 0 ,V(-ysy) > V(-y) ,then E[ ^V 2(-ysy)] < E[ ^V 1(-ysy)] < V(-y) . Where V(-ysy) = 1 K∑k i= 1 珡Y i- 珡Y 2 ,V(-y) = ( 1 M- 1 N) 1 N- 1 ∑K i= 1 ∑M j= 1 Yij - 珡Y 2 ρk = ∑ K i= 1 ∑ M- 1 j= 1( Yij - 珡Y i)( Yi,j+ 1 - 珡Y i) + 1 2 ∑K i= 1( Yi1 - 珡Y i)2 + ∑K i= 1( YiM - 珡Y i)2 ∑K i= 1 ∑M j= 1 Yij - 珡Y i 2
出处
《北京林业大学学报》
CAS
CSCD
北大核心
1999年第5期37-40,共4页
Journal of Beijing Forestry University
关键词
森林
系统抽样
方差估计
相邻单元差值
systematic sampling,estimation of variance,neighbouring unit difference