### Table 6: Equivalent results using stepwise linear regression.

### Table 3.2: Our example information system Remark: It is worth repeating that the Rough Set method does not make any assumption about the nature, interpretation or type of the attributes. In the following, the attributes are described as text strings, but replacing the attribute names with a, b and c, and substituting numbers for the di erent attributes, would yield equivalent results. For this IS, the equivalence classes for an object xi 2 Ei using the attribute set B = fPriceg are:

### Table 1. Persistence time constants for the experimental system used in this work compared with the equivalent results for two com- mercial image intensifier systems

"... In PAGE 2: ... A simple cubic curve-fit was applied in each case and the persistence at which the minimum threshold contrast (corresponding to the maximum signal- to-noise ratio) occurred was recorded. Results The measured time constants corresponding to the weighting factor values W51, 2, 4, 8 and 16 are given in Table1 for the experimental system used here and for two common commercial fluoroscopy systems (data from [3]). There is reasonable correspondence between the time constant values for the various settings, making it possible to interpret the optimum time constant values given below in terms of settings on these units.... In PAGE 4: ... Figure 3 shows the optimum persistence time constant as a function of disk diameter and speed. This can be used in conjunction with Table1 to give a guide to the best temporal averaging selection where the approximate size and speed of the signal is known, e.... In PAGE 4: ...istence at 12.5 mm s21 is constant at a value of 0.15 s for object diameters greater than 4 mm. From Table1 this persistence corresponds to W54, or AVG2 and NR4 in the Philips and GE terminology, respectively. For cardiac motion, the maximum speed studied, 75 mms21 is probably the best guide.... ..."

### Table 2 gives equivalent results for a mildly anisotropic case, with the strong coupling lying at an angle of 23 to the x-axis. h?1

"... In PAGE 7: ... Table2 : Number of iterations (cpu time) [number of restarts] for = 10?3, = 23 . For this problem the spectrum is less clustered, leading to a deterioration in performance of unpreconditioned CG.... ..."

### Table 1: Recognition rates in percentages for the char- acter sets PRC, PRCU, and HWC, using features ex- traced with the cosine, sinus, ordered Hadamard, and the Hartley transforms, and by the method described by Kuhl amp; Gardina, Zahn amp; Roskies, and Lin amp; Hwang. For the set PRCU, all the methods yield equivalent results. On the set of all the printed characters how- ever, we observe that the cosine, sinus and ordered Hadamar features perform better than the Hartley or Fourier-based methods. We were expecting such a re- sult since these three transforms have better conver- gence rates, which implies that an equal number of features retains more informations about the signals. Finally, on hand written characters our methods per- formed much better. This is due to the sampling rate

### Table 1. Verification results for different methods.

2002

"... In PAGE 16: ...Table1 shows the effectiveness of the method we present. The designs in the first column of the table are single compare points that are known as very hard and that existing state-of-the-art equivalence checkers fail to verify with various techniques such as BDD, SAT, ATPG, functional learning and so on.... In PAGE 16: ...able 1. Verification results for different methods. Table 2 shows the performance in terms of time taken and consumed peak memory. In the table, the numbers with (*) show the time and peak memory until the cases were aborted at the depths shown in Table1 . For the designs (D8CTD7D8BGBN D8CTD7D8BHBN D8CTD7D8BJ) in which both NORM and ours can verify, even though ours uses the expensive range computations, the results were faster in D8CTD7D8BH and D8CTD7D8BJ.... ..."

Cited by 10

### Table 2: Recognition performance of various feature representations, Parking Lot data set

2004

"... In PAGE 7: ... 4 Results The various feature extraction methods described in the previous section were tested within the proposed recognition system using the vehicle data described in section 2. Both their direct identification (what type of vehicle is it?) and verification (is it this type of vehicle?) performance is given in Table2 , where Pid is the probability of correct identi- fication and Equal Error Rate, EER the point at which probabilities of false acceptance and false rejection are equivalent. The results were obtained using manually marked number- plate locations, i.... In PAGE 8: ... In general, this data set contains fewer different vehicles (registered vehicles repeat in the test set), however the vehicles exhibit a higher range of out-of-plane rotation and location and the images are of lower quality than the Parking Lot data set. Recognition performance for this data set, given for a chosen subset of structure representations in Table 3, correspond well to the results in Table2 with gradient features giving the best results. This time however, both Harris corner detector and raw pixel values perform well as the number of vehicles in the test set is reduced and colour correspondency exists with the registered examples.... ..."

Cited by 4

### Table 5-4 shows equivalent results for accidents as defined by ICAO. Compared to Table 3-2, there is relatively little change because the small increase in taxiway accidents, which dominate the total, offsets the reduction in other accident causes. There is a general reduction but otherwise little change in the pattern of ATC causes.

2006

"... In PAGE 17: ...able 4-3: Validation Against Trends in Fatal Accident Frequencies since 1990.......................... 48 Table5 -1: Change in Fatal Accident Frequencies to 2012 .... In PAGE 17: ...able 5-1: Change in Fatal Accident Frequencies to 2012 ........................................................... 50 Table5 -2: Change in Collision Accident Frequencies to 2012.... In PAGE 17: ...able 5-2: Change in Collision Accident Frequencies to 2012...................................................... 51 Table5 -3: Contributions of ATC to Direct Causes of Fatal Aircraft Accidents for Commercial Flights within ECAC Region in 2012 .... In PAGE 17: ...lights within ECAC Region in 2012 ............................................................................ 52 Table5 -4: Contributions of ATC to Direct Causes of ICAO-Defined Accidents for Commercial Flights within ECAC Region in 2012 .... In PAGE 70: ...orming a definitive future prediction. Some of the main uncertainties are reviewed in Section 5.3 and used as the basis of sensitivity tests in the results. The predicted overall effects on the accident frequencies of all modelled ATM changes are summarised in Table5 -1. For consistency, the frequencies of accident categories not modelled in IRP are assumed to have continue to reduce at 4.... In PAGE 70: ...5% per year, in line with historical statistics. Table5 -1: Change in Fatal Accident Frequencies to 2012 ACCIDENT CATEGORY 2005 FATAL ACCIDENT FREQUENC Y (per flight) 2012 FATAL ACCIDENT FREQUENC Y (per flight) ACCIDENT FREQUENC Y RATIO 2012/2005 ANNUAL REDUCTIO N RATE (%) Mid-air collision 5.... In PAGE 71: ...076/EEC/NB/05 - EEC Note No. 05/06 51 The contributions of the various ATM changes to the modelled reduction in mid-air collision frequencies are shown in Table5 -2. The main contributors are ATFCM, ACAS and STCA.... In PAGE 71: ... The main contributors are ATFCM, ACAS and STCA. Table5 -2: Change in Collision Accident Frequencies to 2012 ATM CHANGE CHANGE TO 2012 COLLISION FREQUENCY RATIO 2012/2005 Commercial traffic 44% higher 1.20 Military traffic 19% lower 0.... In PAGE 72: ...2. ATM CONTRIBUTION TO ACCIDENT RISKS Table5 -3 shows the contribution of ATC to the direct causes of the five main accident categories in 2012. This can be compared to Table 3-1.... In PAGE 72: ... This result should be treated with caution, as it compares the modelled effects of the ATM changes against the extrapolated statistical reduction of the other causes. Table5 -3: Contributions of ATC to Direct Causes of Fatal Aircraft Accidents for Commercial Flights within ECAC Region in 2012 ACCIDENT CATEGORY FATAL ACCIDENT FREQUENCY (per flight) ATC DIRECT CAUSES (% of category) FREQUENCY OF FATAL ACCIDENT DIRECTLY CAUSED BY ATC (per flight) ATC DIRECT CAUSES (% of total) Mid-air collision 3.1E-09 67.... In PAGE 72: ... There is a general reduction but otherwise little change in the pattern of ATC causes. Table5 -4: Contributions of ATC to Direct Causes of ICAO-Defined Accidents for Commercial Flights within ECAC Region in 2012 ACCIDENT CATEGORY ICAO ACCIDENT FREQUENCY (per flight) ATC DIRECT CAUSES (% of category) FREQUENCY OF ICAO ACCIDENT DIRECTLY CAUSED BY ATC (per flight) ATC DIRECT CAUSES (% of total) Mid-air collision 3.1E-09 67.... ..."

### Table 2 indicates that very similar results can be obtained for a random coe cient second- order elliptic partial di erential equation in two dimensions. Once again the precondi- tioned iterative methods converge in O(h?1=2) iterations. As for Poisson apos;s equation the best preconditioned method is PCGS, which is more than 11 times faster than its un- preconditioned counterpart on the nest mesh. As a direct method F11DCF is a slight improvement over F01BRF/F04AXF for h = 1=32 and h = 1=64, but runs out of storage space on the ner meshes.

"... In PAGE 8: ... Table2 : Number of iterations (CPU time) for a random elliptic PDE. Tables 3 and 4 give equivalent results for the convection-di usion equation ? r2u + a:ru = g (7) with = 10?2 and a = (3; 4)T.... ..."

### Table 7. Comparison of the accuracy of different docking methodsa

2000

"... In PAGE 16: ... The accuracy of the dock results can be compared directly with other methods for which the same test- cases have been studied. Table7 summarizes the RMS distances of the docked positions for the method pre- sented here and for four published methods, including a genetic algorithm method (GOLD5;6) and three in- cremental construction methods. The nearest docked position generated by our method is equivalent to the results of the other methods.... ..."