### Table 1. Application Properties Packet Source Num. of Num. of Distinct

2005

"... In PAGE 6: ... We can exploit this behavior by caching the recently accessed entries in the port mapping table to avoid the lookups. Table1 summarizes the behavior of the applications discussed. Next we present results measuring the degree of value locality in these applications.... In PAGE 6: ... We measured the value locality in terms of the percentage of the packets which have a frequently occurring combination of values for the relevant n- tuple of header fields where n is the number of relevant header fields for a given application. According to Table1 , IP Routing examines a 1-tuple, Packet Classification examines a 5-tuple, and NAT Protocol examines a 4-tuple of fields.... ..."

Cited by 1

### Table 2. Description of the datasets used to measure the computation time (upper table). The column #Distinct indicates the number of distinct values of the property of interest (p.o.i.). Times in seconds and number of rules generated for the datasets for different minimal supports (lower table).

"... In PAGE 7: ... Furthermore our method outputs whole distributions and defines the interest of a rule in terms of comparison of distri- butions rather than the comparison of means. 5 Evaluation In this section we show how our algorithm CAREN-DR performs on 4 different datasets described in Table2 . The algorithm has been run with different values... In PAGE 8: ... We can see that the algorithm scales up quite well with the number of examples and the value of minimal support. Table2 (bottom) shows the times in seconds spent on a Pentium IV, 1.6GHz and 1GB RAM.... In PAGE 8: ...he times in seconds spent on a Pentium IV, 1.6GHz and 1GB RAM. These times include writing the rules to a csv file (one of the possible output modes). Table2 (bottom) also shows the number of rules produced per run. We stress, however, that by turning improvement on, the number of rules falls dramatically.... ..."

### Table 2: Properties of GenBank, with sequences di- vided into intervals (entropy in bits per base, distinct intervals in model).

"... In PAGE 2: ... Now we consider the entropy of our test collec- tions. Results are shown in Table2 , giving the en- tropy Eint n and the number of distinct intervals for each collection and interval length. The entropy is almost exactly as expected for random data.... ..."

### Table 1 summarizes our upper and lower bounds. The rows correspond to the different restrictions on the set A of algorithms, and the columns to the restrictions on the set D of databases and on the aggregation function t. Note that SM means strictly monotone and SMV means strictly monotone in each variable. Distinctness means that D is the collection of databases that satisfy the distinctness property. Note also that c = max {cR cS , cS cR }. For each such combination we provide our upper and lower bounds, along with the theorem where these bounds are proven. The upper bounds are stated above the single separating lines and the lower bounds are below them. (The upper bounds are stated explicitly after the proofs of the referenced theorems.) The lower bounds may be deterministic or probabilistic.

2001

"... In PAGE 37: ...5 access Lower bound: m Thm 9.5 (t strict) Table1 : Summary of Upper and Lower Bounds... In PAGE 39: ...2 says that in the case of no wild guesses and a strict aggregation function, TA is tightly instance optimal. In the case of no wild guesses, for which aggregation functions is TA tightly instance optimal?19 What are the possible optimality ratios? For the other cases where we showed instance optimality of one of our algorithms (as shown in Table1 ), is the algorithm in question in fact tightly instance optimal? For cases where our algorithms might turn out not to be tightly instance optimal, what other algorithms are tightly instance optimal? There are several other interesting lines of investigation. One is to find other scenarios where in- stance optimality can yield meaningful results.... ..."

Cited by 231

### Table 4: Property usage in FOAF dataset

2005

"... In PAGE 9: ... In order to evaluate their utility in practice, we collected statistics about the properties being used to de- scribe instances of foaf:Person. We found 546 distinct properties used for at least one Person instance, as shown in Table4 . Only 34 properties were used by more than 1% of the FOAF documents.... ..."

Cited by 2

### Table 1. Properties of the two sets of images used in this study. A set of 12 cases was used for observer training and measurements were collected using a second, distinct set of 21 cases

"... In PAGE 2: ... A set of 12 cases was used for observer training and measurements were collected using a second, distinct set of 21 cases. The characteristics of the measurement set are summarized in Table1 . A list of the DDSM cases numbers and the ROI images used in this study are available on our website: http://www.... ..."

### Table 5.2. Given these two graph types we will analyze two types of properties: (i) structural and (ii) dynamical. The former captures the structure of social relationships between users exchanging emails, while the latter relates to how graphical properties evolve over time. As we shall show below there are distinct independent signatures of spam traf c in both structural and dynamical properties. As a consequence they should be taken together to generate a better detection procedure.

2006

### Table 3: Number of Distinct Prefix Lengths in the 16 bit Partitions (Histogram)

"... In PAGE 7: ...Table 3: Number of Distinct Prefix Lengths in the 16 bit Partitions (Histogram) nice property of the routing data ( Table3 ). Though the whole his- togram (Table 7) shows 23 distinct prefix lengths with many buck- ets containing a significant number of entries, none of the sliced histograms contain more than 12 distinct prefixes; in fact, the vast majority only contain one prefix, which often happens to be in the 16 bit prefix length hash table itself.... ..."

### Table 3: Number of Distinct Prefix Lengths in the 16 bit Partitions (Histogram)

"... In PAGE 7: ...Table 3: Number of Distinct Prefix Lengths in the 16 bit Partitions (Histogram) nice property of the routing data ( Table3 ). Though the whole his- togram (Table 7) shows 23 distinct prefix lengths with many buck- ets containing a significant number of entries, none of the sliced histograms contain more than 12 distinct prefixes; in fact, the vast majority only contain one prefix, which often happens to be in the 16 bit prefix length hash table itself.... ..."

### Table 3: Number of Distinct Prefix Lengths in the 16 bit Partitions (Histogram)

"... In PAGE 7: ...Table 3: Number of Distinct Prefix Lengths in the 16 bit Partitions (Histogram) nice property of the routing data ( Table3 ). Though the whole his- togram (Table 7) shows 23 distinct prefix lengths with many buck- ets containing a significant number of entries, none of the sliced histograms contain more than 12 distinct prefixes; in fact, the vast majority only contain one prefix, which often happens to be in the 16 bit prefix length hash table itself.... ..."