### Table 1 Polymorphically order-sorted unification.

"... In PAGE 6: ... As in the polymorphic cases above, we now have not only a finitc set of sorts but infinitely many sorts denoted by sort te~ which show up in the unification procedure: Besides a sort constant such as car or vehicle, the sort restriction of a variable can be a sort term such as list(car) or list(list(vehic1e)). To illustrate polymorphically order-sor unification, we list in Table1 the unification results for different values of the terms tl and t2. The partial order on the (monomorphic) sort constanl induces a partial order on sort terms as shown in the ta For instance, list(amphibious-vehicle) is a subsort of list(car) because amphibious-vehicle is a subsort of car.... ..."

### Table 1: Run times for k-k sorting in row-major or- der. Lower-order terms are left away in the results for large n. The queue sizes in the algorithms for 1-1 and 2-2 sort- ing range between four and nine, and in the k-k sort- ing from k to k + 2. Actually the result for large k and n is much more general: it does not just hold for sorting in row-major order, but for sorting with respect to any indexing that is piecewise-continuous (see De nition 1 in Section 2). Theoretically the result for large n are the most ap-

1994

"... In PAGE 16: ...76 9 210 2 3 5 7 942 4.49 9 Table1 0: Actual numbers of steps required by sort- all for uni-axial row-major sorting. In the version of [10], the mesh is divided in s s submeshes with s = n2=3=k1=3, and the randomiza- tion of Step 1 is replaced by sorting the packets in the submeshes and unshu ing them regularly over the submeshes.... ..."

Cited by 17

### Table 5: Combining reflected-order sorting with alphabet reordering.

1998

"... In PAGE 8: ...hange occurred for j gt;8. This is summarized in Table 4. The reflected ordering can of course be combined with the alphabet reordering as described in the two preceding sections. Table5 shows the results of combining reflection (using maximum number of columns) with some of the sort orders from previous sections. 3 Best overall results Table 6 takes the best compression results from previous tables and shows the methods used on each le to achieve the result.... ..."

Cited by 10

### Table 1: Confusion matrices illustrating classification performance using single and all terms in the distance metric.

2006

"... In PAGE 9: ... In this case we may need to employ additional features, based on appearance or shape, for example, in order to sort out the ambiguity. The confusion matrix illustrating the classification perfor- mance is shown on the right in Table1 . Also shown are the confusion matrices resulting when only one of the terms in the distance metric is used, i.... ..."

### Table 1: Confusion matrices illustrating classification performance using single and all terms in the distance metric.

"... In PAGE 8: ... In this case we may need to employ additional features, based on appearance or shape, for example, in order to sort out the ambiguity. The confusion matrix illustrating the classification perfor- mance is shown on the right in Table1 . Also shown are the confusion matrices resulting when only one of the terms in the distance metric is used, i.... ..."

### Table 10 provides the results of this approach for the selection model. The hospitals listed in the table are sorted by the values of the r j , and the number to the left of each hospital name indicates its order in this sorting. The r j are shown in the second column following the hospital names. Posterior mean quality is shown in the first column following the names. Ordering by hospital quality posterior mean does not lead to the same ordering as ordering by r j , but it is very

1999

"... In PAGE 27: ...the posterior distribution of rj , which in turn is an integer (with probability one). Medians are shown in the third column of Table10 . They provide yet another ordering, but it too is similar to ordering by mean quality and mean rank.... In PAGE 27: ... Placement within a quartile is most certain for hospitals of very high or very low quality. For those hospitals ordered 20 through 102 in Table10 , the posterior probability is at least .10 that the hospital is in one of each of three quartiles.... In PAGE 27: ... At the other extreme, if hospital qualities were completely exchangeable in the posterior distribution, the mean and median ranks would all be 59. Note that the situation in Table10 is intermediate between these two extremes, but closer to the former than the latter. Table 11 provides an alternative expression of the uncertainty conveyed by the posterior distribution.... In PAGE 27: ... Table 11 provides an alternative expression of the uncertainty conveyed by the posterior distribution. Nine hospitals, including the first and last, were selected from roughly evenly spaced points in the ordering in Table10 . Then, pairwise posterior probabilities of orderings were computed from the iterations of the Gibbs sampler.... In PAGE 28: ... From this perspective, the greater posterior confidence that is seemingly inherent in the probit model is ironic. Note that the information in Table10 cannot be used to address the question of the range of quality over all 117 hospitals. The posterior mean of the highest quality probit of all hospitals... ..."

Cited by 2

### Table 10 provides the results of this approach for the selection model. The hospitals listed in the table are sorted by the values of the r j , and the number to the left of each hospital name indicates its order in this sorting. The r j are shown in the second column following the hospital names. Posterior mean quality is shown in the first column following the names. Ordering by hospital quality posterior mean does not lead to the same ordering as ordering by r j , but it is very

1999

"... In PAGE 27: ...the posterior distribution of rj , which in turn is an integer (with probability one). Medians are shown in the third column of Table10 . They provide yet another ordering, but it too is similar to ordering by mean quality and mean rank.... In PAGE 27: ... Placement within a quartile is most certain for hospitals of very high or very low quality. For those hospitals ordered 20 through 102 in Table10 , the posterior probability is at least .10 that the hospital is in one of each of three quartiles.... In PAGE 27: ... At the other extreme, if hospital qualities were completely exchangeable in the posterior distribution, the mean and median ranks would all be 59. Note that the situation in Table10 is intermediate between these two extremes, but closer to the former than the latter. Table 11 provides an alternative expression of the uncertainty conveyed by the posterior distribution.... In PAGE 27: ... Table 11 provides an alternative expression of the uncertainty conveyed by the posterior distribution. Nine hospitals, including the first and last, were selected from roughly evenly spaced points in the ordering in Table10 . Then, pairwise posterior probabilities of orderings were computed from the iterations of the Gibbs sampler.... In PAGE 28: ... From this perspective, the greater posterior confidence that is seemingly inherent in the probit model is ironic. Note that the information in Table10 cannot be used to address the question of the range of quality over all 117 hospitals. The posterior mean of the highest quality probit of all hospitals... ..."

Cited by 2

### Table 10 provides the results of this approach for the selection model. The hospitals listed in the table are sorted by the values of the r j , and the number to the left of each hospital name indicates its order in this sorting. The r j are shown in the second column following the hospital names. Posterior mean quality is shown in the first column following the names. Ordering by hospital quality posterior mean does not lead to the same ordering as ordering by r j , but it is very

1999

"... In PAGE 27: ...the posterior distribution of rj , which in turn is an integer (with probability one). Medians are shown in the third column of Table10 . They provide yet another ordering, but it too is similar to ordering by mean quality and mean rank.... In PAGE 27: ... Placement within a quartile is most certain for hospitals of very high or very low quality. For those hospitals ordered 20 through 102 in Table10 , the posterior probability is at least .10 that the hospital is in one of each of three quartiles.... In PAGE 27: ... At the other extreme, if hospital qualities were completely exchangeable in the posterior distribution, the mean and median ranks would all be 59. Note that the situation in Table10 is intermediate between these two extremes, but closer to the former than the latter. Table 11 provides an alternative expression of the uncertainty conveyed by the posterior distribution.... In PAGE 27: ... Table 11 provides an alternative expression of the uncertainty conveyed by the posterior distribution. Nine hospitals, including the first and last, were selected from roughly evenly spaced points in the ordering in Table10 . Then, pairwise posterior probabilities of orderings were computed from the iterations of the Gibbs sampler.... In PAGE 28: ... From this perspective, the greater posterior confidence that is seemingly inherent in the probit model is ironic. Note that the information in Table10 cannot be used to address the question of the range of quality over all 117 hospitals. The posterior mean of the highest quality probit of all hospitals... ..."

Cited by 2

### Table 3.1. Bit order in generalized Peano-order sorting.

### Table 4: The top 20 EPs and the top 20 strong EPs, in a descending order, sorted by their frequency in the 40 cancerous tissues.

2001

"... In PAGE 5: ... And a total of 2165 EPs, which have a non-zero frequency in the cancerous tissues, were derived by our algorithms. According to the frequency, Table 3 (presented at the last page) and Table4 list the top 20 EPs and strong EPs which occur in the 22 normal tissues, and the top 20 EPs and strong EPs which occur in the 40 cancerous tissues. Column 1 shows the emerging patterns.... In PAGE 8: ...lass, but never contain any EPs from the other class. So, our system has well learned the whole data. Next we use emerging patterns to perform a classification task to see how useful of the patterns in predicting whether a new cell is normal or cancerous. 6 The Usefulness of EPs in Classification As shown in Table 3 and Table4 , the frequency of the EPs is very large. Obviously, the groups of genes are good indicators for classifying new tissues.... ..."

Cited by 7