Figure 6.66. MPLS-based on-line monitoring results of a faulty batch.
(r0jt) at time k and the recent history (taf(rc_1j), playing a role similar to that of the exponential weighting factor in an EWMA model. The consensus matrix R0fc is formed from ta(k~i) and rak column vectors and the weight vector wa/- is computed as (wafc = Rjfctafc) for calculating the new score vector tak = Rakwak- Then, tak is normalized and checked for convergence.
If convergence is achieved the Xafc blocks are deflated as X(a.
Was this article helpful?