### Table 1. Values of t for the (t; m; s)-Nets NET Parameters

1998

Cited by 17

### Table 1: Upper bounds on s for existence of (t; m; s)-nets in base 2

"... In PAGE 11: ... Applying Lemma 4.3, we have the following more e cient way of computing S (t; m; b): S (t; m; b) = maxfs : bn GR(n?t; s; n?t; b) for n?t even, t+2 n mg: We tabulate S (t; m; 2) for 1 t 11, t + 2 m 15, m ? t even, in Table1 , where we also record the best previously known upper bounds on s from [1] (as given in the Tables on the web site http://www.... In PAGE 11: ... (As mentioned before, we can restrict our attention to m?t odd since S (t; m; b) = S (t; m?1; b) if m?t is odd.) For the values of m and t considered in Table1 , our new bound is at least as good as the bound from [1], except for (t; m) = (2; 6) and (5; 13). There are numerous cases where our bound improves the bound from [1].... In PAGE 12: ... In a similar manner, we tabulate S (t; m; 3) for 1 t 11, t + 2 m 15, m ? t even, in Table 2, comparing it to the best previously known bounds upper bounds on s from [1]. As was the case in Table1 , we nd a signi cant number of improvements. 6 Comments We have derived a generalized Rao bound that can be applied to ordered orthogonal arrays and (t; m; s)-nets.... ..."

### Table 2: Upper bounds on s for existence of (t; m; s)-nets in base 3

"... In PAGE 12: ... Our bound is the same as the Mullen-Whittle bound when m = t + 2 or m = t + 3; it is easily veri ed that S (t; t + 2; b) = S (t; t + 3; b) = bt+2 ? 1 b ? 1 : In these cases, it is also often the case that the bound is tight; see [4] for more details. In a similar manner, we tabulate S (t; m; 3) for 1 t 11, t + 2 m 15, m ? t even, in Table2 , comparing it to the best previously known bounds upper bounds on s from [1]. As was the case in Table 1, we nd a signi cant number of improvements.... ..."

### Table 2. New lower bounds on T for binary (T; M; S)-nets Lower Bound Pairs (M; S) with reference to LP bounds T 3

1998

Cited by 4

### Table 1: Upper bounds on s for existence of (t; m; s)-nets in base 2 t m S (t; m; 2) [1] t m S (t; m; 2) [1]

"... In PAGE 11: ... Applying Lemma 4.3, we have the following more e cient way of computing S (t; m; b): S (t; m; b) = maxfs : bn GR(n?t; s; n?t; b) for n?t even, t+2 n mg: We tabulate S (t; m; 2) for 1 t 11, t + 2 m 15, m ? t even, in Table1 , where we also record the best previously known upper bounds on s from [1] (as given in the Tables on the web site http://www.... In PAGE 12: ...Table1 , our new bound is at least as good as the bound from [1], except for (t; m) = (2; 6) and (5; 13). There are numerous cases where our bound improves the bound from [1].... In PAGE 12: ... In a similar manner, we tabulate S (t; m; 3) for 1 t 11, t + 2 m 15, m ? t even, in Table 2, comparing it to the best previously known bounds upper bounds on s from [1]. As was the case in Table1 , we nd a signi cant number of improvements. 6 Comments We have derived a generalized Rao bound that can be applied to ordered orthogonal arrays and (t; m; s)-nets.... ..."

### TABLE VI THE AVERAGE ISCHEMIA EPISODE DETECTION PERFORMANCE EVALUATED WITH THE CORRESPONDING NETWORKS (i.e., SOM, sNet-SOM WITH RBF AS SUPERVISED EXPERT AND sNet-SOM WITH SVM AS SUPERVISED EXPERT

2001

Cited by 5

### Table 2: Upper bounds on s for existence of (t; m; s)-nets in base 3 t m S (t; m; 3) [1] t m S (t; m; 3) [1]

"... In PAGE 12: ... Our bound is the same as the Mullen-Whittle bound when m = t + 2 or m = t + 3; it is easily veri ed that S (t; t + 2; b) = S (t; t + 3; b) = bt+2 ? 1 b ? 1 : In these cases, it is also often the case that the bound is tight; see [4] for more details. In a similar manner, we tabulate S (t; m; 3) for 1 t 11, t + 2 m 15, m ? t even, in Table2 , comparing it to the best previously known bounds upper bounds on s from [1]. As was the case in Table 1, we nd a signi cant number of improvements.... ..."

### TABLE 1: Algorithm to construct a constraint apos;s net Until stop by user

### Table 5: Co-selection of sentences between the program and BT apos;s NetSumm.

### Table 1 The worst and the best e ciencies, that is, En0;n0+n 1 and En0;2n0, among the sequences obtained by scrambling unions of (0; m; s)-net and ( 1; 0; m ? d; s)-net in base b with 1 1 lt; b, 0 d m, where if d = 1 then 1 = 1.

"... In PAGE 14: ... Figure 3 shows the values of En0;n for n0 = bm and n = n0 + 1bm?d with 1 1 lt; b, 1 d m, using b = 7 and 11, m = 5, 6 and 7 respectively. Table1 lists the \worst quot; and the \best quot; e ciencies among sequences of the form (0; m; s) [ ( 1; 0; m; s)-net in base b with 1 1 lt; b, 0 d m, where if d = 0 then 1 = 1. These results show that among sequences used with the number ranged from 0bm to ( 0 + 1)bm of points, the one with ( 0 + 1)bm points, that is ( 0 + 1; 0; m; s)-net, has the smallest variance.... ..."