A Control Chart Based on Cluster-Regression Adjustment for Retrospective Monitoring of Individual Characteristics

Abstract

The tendency for experimental and industrial variables to include a certain proportion of outliers has become a rule rather than an exception. These clusters of outliers, if left undetected, have the capability to distort the mean and the covariance matrix of the Hotelling’s T 2 multivariate control charts constructed to monitor individual quality characteristics. The effect of this distortion is that the control chart constructed from it becomes unreliable as it exhibits masking and swamping, a phenomenon in which an out-of-control process is erroneously declared as an in-control process or an in-control process is erroneously declared as out-of-control process. To handle these problems, this article proposes a control chart that is based on cluster-regression adjustment for retrospective monitoring of individual quality characteristics in a multivariate setting. The performance of the proposed method is investigated through Monte Carlo simulation experiments and historical datasets. Results obtained indicate that the proposed method is an improvement over the state-of-art methods in terms of outlier detection as well as keeping masking and swamping rate under control.

طراحی نمودار کنترل T2 هتلینگ با استفاده از خوشه‌بندی

“…`} RwD |=v@tQ@ = Q 1 R=i |rQDvm OwOL [9] nv=} w uwU}t 'CU= n = 1 xm |Dr=L QO %Ov=xDiQo Q_v QO T 2 pQDvm Q=Owtv |= Q@ 4 |x]@= Q CQwYx@ 8 < : UCL = p(m 1) 2 m ; p/2; m p 1/2 LCL = 0 (4) [10] "CU= ?U=vt 'OW=@ Vy=m |yHwD p@=k Qw] x@ =yl}vmD u}= CQOk 'OW=@ xDW=O OwHw =yxO=O xawtHt xDW=O OwHw Ov}=Qi QO CQB |xO=O l} R= V}@ xm |t=ovy |DQ=@a x@ [12w11] "O@=}|t [13] "OwW|t pmWt CQB \=kv X}NWD 2 '|Q@uwQ@ w 1 |Qw;uwQO QF= p}rO x@ 'OW=@ CQB u=wvax@ |ak=w CQB C=Oy=Wt X}NWD xm ODi=|t j=kD= |t=ovy |Qw;uwQO CQwYx@ C=Oy=Wt xm ODi=|t j=iD= |t=ovy |Q@uwQ@ xm |r=L QO "OwW|tv sqa= C}r@=k 'hWm sOa CQwY QO 'CQB |=yxO=O [14] "OwW sqa= CQB u=wvax@ CUQO=v nv}rDy T 2 |xQ}eDtOvJ pQDvm |=yQ=Owtv Tv=} Q=wwm T} QD=t w u}ov=}t h= QLv= xm CU= u}= h} QLD u}= Q}F -=D "OvQ=O = Q |OQi C}i}m |=y|oS}w Q@ CQ=_v |= Q@ w |Qw;uwQO QF= |= Q=O = Q} R 'OwW|t O=tDa= p@=kQ}e u; R= xOW xDN=U pQDvm Q=Owtv Q=}U@ =yxO=O p}rLD w x} RHD R= p@k CQB \=kv |}=U=vW 'u}= Q@=v@ [10] "CU= |Q@uwQ@ [8] "O}rwD Ov}=Qi pwYLt C}i}m Q@ CQ=_v QO XwYNx@ CU= syt |R=Ux}@W |=yl}vmD R= xO=iDU= =@ T 2 |}xU}=kt |xar=]t l} QO [22] T=oQ=w XN=W R= xO=iDU= =@ w CQBxO=O l} X}NWD T=U= Q@ = Q |rQDvm Q=Owtv OQmrta O=O u=Wv OQmrta |@=} RQ= w xrY=L G}=Dv "OQm |UQQ@ (ncp) C} RmQt sOa (MVE) |[}@ sHL pk=OL |=yQoOQw; Q@ R= xO=iDU= =@ T 2 pQDvm Q=Owtv xm O=OaD |}=U=vW QO SW1 VwQ w CU= QFwt CQB \=kv O=OaD |}=U=vW QO =@ |}=yxO=O p@=kt QO R}v SW2 VwQ "CU}v QFwt |i=QLv= \=kv R= |O=}R xDU@=w x}rw= |iO=YD |xvwtv x@ u}vJty w CU= Q}PB? }U; 'O=}R xO=Di= QwO \=kv "CU= VwQ R= xO=iDU= =@ T 2 |=H x@ C=@F=@ QoOQw;Q@ l} [16] =oDQ= w wQ=ir= Q=OQwNQ@ T 2 x@=Wt Q=DN=U R= O}OH pQDvm Q=Owtv u}= "OvOQm O=yvW}B xDU=Q}B x@ V}=Q}B R= xO=iDU= =@ Tv=} Q=wwm T} QD=t OQw; Q@ w C}akwt Q=OQ@ =t= 'CU= p=Ft l} R= xO=iDU= =@ = Q Q=Owtv wO u}= OQmrta w Q=DiQ =yu; "O};|t CUO xU}=kt CQB |=yxO=O |xawtHt X}NWD T=U= Q@ |R=Ux}@W xar=]t l} w p}rLD |= Q@ |R=Ux}@W |xar=]t l} [11] QoVywSB wO u}= Oa@ p=U l} "OvOQm QO = Q xDU=Q}B QoQw;Q@ w uwRwtR=@ MCD 'MCD 'MVE |=yQoOQw; Q@ OQmrta ?…”

“…""" nv}rDy T 2 pQDvm Q=Owtv |L=Q] Tv=} Q=w T} QD=t w u}ov=}t uwJ ODi=|t j=iD= |Q@uwQ@ w |Qw;uwQO |xO}OB R= xO=iDU= 'CqmWt u}= pL |=yx= Q R= |m} [10] "OvDU}v C=@F=@ Tv=} Q=wwm w T} QD=t w C}akwt |= Q@ |}xv} Ro u=wvax@ 3 w 2 CqO=at |= Q@ C=@F=@ |=yQoOQw; Q@ [16w15] 'nv}rDy T 2 Ow@y@ =@ 3 xDU=Q}B VwQ pt=W =yQoOQw;Q@ u}= "CU= Tv=} Q=wwm KqY= w OvwW|t xO}t=v SW1 xm CU= [17] p=Oww w u=w}r=U |B QO|B hqDN= VwQ xm 'p=Oww w u=w}r=U swO OQm}wQ u=wvax@ [18] = Q}rwt w uwUv}mD= 4 C}Dm q=DU= KQ] Ow@y@ = Q nv}rDy T 2 Ovv=wD|t `k=w QO xOW Qm P |=yVwQ "Ov}wo|t SW2 = Q u; QFwt =yxO=O QO OwHwt CQB \=kv R= |O=}R O=OaD X}NWD QO Rwvy =t= 'OvWN@ [8] "OvDU}v VwQ pk=OL w 5 (MVE) Q=w|[}@ sHL pk=OL VwQ wO [19] wUQ |O=}R O=OaD X}NWD |xrUt |= Q@ = Q 6 (MCD) Tv=} Q=wwm |xOvvmu}}aD Q=}U@ C=@U=Lt x@ R=}v VwQ wO Qy 'p=L u}= =@ "OvOQm O=yvW}B CQB \=kv R= MVE ?} QkD OQwt QO |Q=}U@ |=yEL@ xDWPo |xyO wO QO "OQ=O u}ovU xO=iDU= C=ar=]t R= |NQ@ QO 'Q}N= |xyO QO =t= "CU= xOW s=Hv= MCD w "CU= xOW R=e; xQ}eDtOvJ pQDvm Q=Owtv Ow@y@ |= Q@ |}xWwNp}rLD |=yVwQ R= R= [20] s}m s=@ w nv=m 'uw}UQoQ |}xWwN p}rLD R= [10] |r= w nvw= p=Ft |= Q@ |@D=QtxrUrU |}xWwN p}rLD R= [8] Vv=Q=mty w u=i 'k-means |}xWwN p}rLD |Ov@xWwN |=yOQm}wQ "Ov=xOQm xO=iDU= SW2 xQ}eDtOvJ pQDvm Q=Owtv |= Q@ =@ =Pr [8] "CU= pQDvm Q=Owtv |=yVwQ R= QDOt;Q=m |D=@U=Lt ^=Lr R= , qwtat QO [8] 'CU= T=UL CQB |=yxO=O x@ C@Uv |}xWwN p}rLD xm u}= x@ xHwD R=i QO pQDvm Q=Owtv |L=Q] QO |@D=QtxUrU |Ov@xWwN l}vmD R= VywSB u}= w OwW|t xO=iDU= C=Oy=Wt R= C=@F=@ |}xawtHt uOQw; CUO x@ |= Q@ 'pw= CQB |=yxO=O |x}rw= QDr}i u=wvax@ |}xWwN |=yOQm}wQ R= xO=iDU= Q@ RmQtD "CU= CQB |xO=O u}OvJ uO=O MQ 'Q=Owtv uOW pQDvm R= GQ=N G}=Q pt=wa R= |m} C@Uv 'nv}rDy T 2 pQDvm Q=Owtv QO , =YwYNt P w |=yQDt= Q=B OQw; Q@ "CU= pQDvm Q=Owtv R= xOt; CUO x@ |=yxQ=t; =t= Ov=T=UL CQB |xOy=Wt l} x@ 'OW=@ xDW=O OwHw CQB xOy=Wt u}OvJ xm |Dkw =yQoOQw;Q@ u}= T=U= Q@ 'QwmPt |= Q@ |O=}R |=yVywSB QO OW xDio xm Qw]v=ty [21] "Ovm|t pta h}a[ Q=}U@ O=OaD X}NWD QO Rwvy =t= 'CU= xOW s=Ok= 'nv}rDy T 2 pQDvm Q=Owtv Ow@y@ p}rLD R= xO=iDU= p}rO =Pr [8] "OvDU}v QFwt =yxO=O QO OwHwt CQB \=kv R= |O=}R u=mt= |@D=QtxrUrU |Ov@xWwN xm CU= u}= Q[=L j}kLD QO |@D=QtxrUrU |}xWwN =yxO=O |t=tD uwJ u}vJty "OQ=O = Q =yxO=O u}@ QO hrDNt |=yxrY=i R= xO=iDU= sy=Qi hrDNt |=ywor= =@ CQB \=kv |}=U=vW u=mt= 'OvwW|t xU}=kt Qo}Om} =@ [10] "OwW|t l} X}NWD T=U= Q@ =OD@= |O=yvW}B pQDvm Q=Owtv OQmrta |@=} RQ= |= Q@ "OwW|t xO=iDU= [22] T=oQ=w C} RmQt sOa XN=W R= xO=iDU= =@ w CQB |xO=O R= GQ=N u}ov=}t Q=OQ@ Q}}eD COW |xOvyOu=Wv (ncp) C} RmQt sOa XN=W "CU= 1 |x]@= Q j@=]t ( 0 ) u}ov=}t pQDvm CLD u}ov=}t Q=OQ@ R= ( 1 ) pQDvm xDUv=O =yxO=O QO p=vo}U l} smCUO X}NWD = Q VwQ CkO Q=}at T=oQ=w "CU= ncp = ( 1 0 ) 0 X 1 ( 1 0 ) (5) T=U= Q@ OQmrta |@=} RQ= R= 'T=oQ=w |@=} RQ= VwQ Q@ xwqa j}kLD u}= QO u}= QO "OwW|t xO=iDU= R}v [11] =oDQ= w wQ=ir= CQB |...…”

An attribute control chart for multivariate Poisson distribution using multiple dependent state repetitive sampling

“…The attribute control charts are relatively easy to apply as compared with variable control charts. As mentioned by Bersimis, “expensive and time‐consuming measurements may be avoided by attributes inspection.” More details and application of control charts can be seen in previous studies …”

“…More details and application of control charts can be seen in previous studies. [2][3][4][5][6][7][8][9] The c and u charts have been widely used in the industry to monitor the number of nonconforming items for the univariate process. However, monitoring more than one quality characteristic is also required in many industries such as telecommunication and electrolysis processes.…”

An Efficient Shewhart-Type Control Chart to Monitor Moderate Size Shifts in the Process Mean in Phase II

“…Abbassi and Riaz suggested control chart for Phase I. Aksioma et al proposed the use of multivariate regression adjustment control charts to determine the mean and variability of the production process. Ong and Alih proposed a control chart that is based on cluster‐regression adjustment for monitoring individual characteristics. Abbassi and Riaz proposed in SPC by making dual use of auxiliary information.…”