### Table1: for Non-Stationary CPTs Time Parents Combination 6 8 11

"... In PAGE 14: ... Thus, despite the fact that an action is still in effect, it may loose its significance as time passes by. The non- stationary CPT values used in the above computations are compared in Table1 along with the time of their computation. Table1: for Non-Stationary CPTs Time Parents Combination 6 8 11 ... ..."

### Table1: for Non-Stationary CPTs Time Parents Combination 6 8 11

"... In PAGE 13: ... Thus, despite the fact that an action is still in effect, it may loose its significance as time passes by. The non- stationary CPT values used in the above computations are compared in Table1 along with the time of their computation. Table1: for Non-Stationary CPTs Time Parents Combination 6 8 11 ... ..."

### Table 3.1: Model parameters 3.2 Expressiveness of the Update Model The proposed update model can generate both stationary and non-stationary updates. In general, when one sets WI lt; S or WD lt; S, the update model generates a non-stationary update process. If one sets WD = WI = S, the update model instead generates a stationary, random update process. Table 3.2 shows the model settings that can be used to generate types of update streams (of length rtot) that have been considered in previous evaluations of incremental histogram maintenance techniques. Stream Type WI WD rinit r L

### Table 2: 5 APPLICATION TO NON-STATIONARY POISSON PROCESSES

1998

Cited by 2

### Table 3: Non-Stationary Exponential Distribution Experiment: Exponential Cases

### Table 8. FPGA resource utilization for uni-variate non-stationary growth model implementation.

"... In PAGE 15: ...ystem.............................................................................................................. 80 Table8 .FPGA resource utilization for uni-variate non-stationary growth model implementation.... ..."

### Table 4. Performance of SFQ with stationary and non-stationary best-effort arrivals.

1999

"... In PAGE 6: ... We effectively induce a total of ten peak bursty arrival peri- ods throughout the duration of the simulation. Table4 illus- trates the performance measures for experiments 4 to 6. The non-stationary process has no effect on the missed deadlines for Canyon.... In PAGE 6: ... The non-stationary process has no effect on the missed deadlines for Canyon. The slowdown of both short and long jobs be- comes higher than the one shown in Table4... In PAGE 9: ...tributed. A-SFQ behaves almost identically to SFQ with the Canyon workload (see Table4 ). The multimedia demand is so low that the work-conserving behavior of SFQ is enough for the system to balance the CPU proportions among the two application classes.... ..."

Cited by 10

### Table 1: Recursive least L normed errors training algorithm. If the lter is to be used in a non{stationary signal environment, the partition H may need to be periodically updated. To account for the changing signal statistics an exponential \forgetting quot; factor can easily be added to the sum of L normed estimate errors. The sum to be minimized is now

1994

"... In PAGE 20: ... The sum to be minimized is now E (M; H) = M X n=1 (M?n)jd(n) ? FP(x(n); H)j ; (42) where 2 (0; 1] is the \forgetting quot; factor. The recursive permutation lter training algorithm that minimizes E (M; H) in (42) is sum- marized in Table1 . The algorithm initially sets the permutation lter to the median lter, and then updates the decision vector according to each new observation.... ..."

Cited by 5

### Table 5. Performance of A-SFQ under stationary and non-stationary best-effort arrivals.

1999

"... In PAGE 8: ... 4.2 Performance of A-SFQ under Stationary ar- rivals of Best-effort Jobs Table5 presents the performance measures for experi- ments 1 to 3.... ..."

Cited by 10