### Table 3. Promising approaches

"... In PAGE 4: ...ng data skewness. If one step makes little difference (e.g., feature selection for DT), we just set No as default to save computation time. Table3 lists the 12 promising ap- proaches to tackle data skewness. The approaches in Table 3 are derived from bias analysis.... In PAGE 4: ... Table 3 lists the 12 promising ap- proaches to tackle data skewness. The approaches in Table3 are derived from bias analysis. We now further evaluate them through comparative experi- ments to investigate whether they can improve performance of classifiers for text classification, and which one is more appropriate for highly skewed data.... ..."

### Table 2. The 12 best collocations of type V-N obtained. The results clearly show that the combination of an accurate parsing and the use of Log-Likelihood ratio leads to a promising approach. When unable to create a complete analysis of a sentence, the Fips parser returns chunks of partial analyses. If

2003

Cited by 2

### Table 4.7: Results of tolerance 2 MCC-cut on data with false positives and 0% false negatives Our last attempt at trying to deal with these vexing false positives is to use LBFS and a method which more explicitly attempts to get rid of the unwanted edges, rather than solely relying on ordering properties of the search. The most promising approach is based on an assumption about the relative frequency of YACs with false positives. If we can assume that the number of vertices on a false branch of any branch point is usually less than the 26

1997

### Table 3 summarizes the di erent forms of predicates de ned in Table 1, and shows the formulae for estimating their selectivity. In this table, a is an element of the argument set A, c is some constant not computed from a taken from some set C, and C is some set not computed from elements of A. The rst column in the table shows the form of the predicate, the second column the selectivity, the third column an approximately equal selectivity derived using the uniform distribution assumption, and the fourth column the resulting estimation formula using one of the equations from the previous section. Experiments with the derived formulas, reported in detail in [NCN95] show the promise of this approach. 7The complete derivations can be found in [NCN95].

1997

"... In PAGE 9: ...Selectivity Equivalent Selectivity Estimation Formula a:~f = c sel(a:~f = c; A) sel(a0 = c; A:~f) 1 jSKCS(A:~f;C)j a:~f 2 C sel(a:~f 2 C; A) sel(a0 2 C; A:~f) jCj jSKCS(A:~f;C)j a:~f = a:~g sel(a:~f = a:~g; A) sel(a0 = a00; A:~f A:~g) 1 jSKCS(A:~f;A:~g)j a:~f 2 a:~g sel(a:~f 2 a:~g; A) sel(a0 2 a00; A:~f A:~g) jA:~gj jSKCS(A:~f;A:~g)j Table3 : Selectivity estimation formulae. bounds on the domain of the attributes of the elements in a set, giving a better estimate for selectivity.... ..."

Cited by 1

### Table 10: The execution time ratio after the two delay optimizations and loop merging the cost of the control structures cannot be neglected when small collections are used. A solution to this problem could be the development of a new control scheme based upon an automaton where each state would represent a di erent clock state con guration (in the way of temporal synchronous languages such as Lustre [2] and Signal [3]). But, because of 81/2 apos;s ability to mix di erent clocks, we should expect rather large automata. The current execution scheme corresponds to a sequential control ow. A promising approach for a better exploitation of the new parallel architectures would be to exploit a parallel control ow. The 81/2 team has already worked in this way, particularly in the mapping and scheduling problems [53] that will be integrated in the present compiler.

### Table 4: Throughput of both hand-coded and ESTEREL versions On the code size aspect, the Protocol Compiler produces a code which has quite the same size as the hand-code written in C language. The executable code of the receiver side is even smaller (4% on 200 Kbytes). Recent optimization work on the ESTEREL apos;s code generation phase show that a sensible reduction of the generated code size is possible without a loss of performance. This is very promising for our automated approach.

1995

Cited by 14

### Table 2 shows probabilities of success for the best function for several cases. Since ^ ws has a carry-save representation, di erent number of bits can be used from the sum and from the carry. When sum and carry bits are expressed in parenthesis, (s-c), the cor- responding bits are rst assimilated and then used by the speculation function. Since the probability of suc- cess is relatively high and the number of bits signif- icantly smaller than those of Table 1, the approach looks promising.

1993

"... In PAGE 5: ... Table2 : Probability of success of the quotient-digit speculation function. 3.... ..."

Cited by 5

### Table 4 shows the execution time for each Replayer step for when only the cookie check is enabled (same as in Table 3), for when the checksum is computed over 16 bytes of input, and for when the checksum is computed over 80 bytes of input. As one would expect, the checksum computation signi cantly increases the complexity of the generated formula. Replayer handles each of these cases in a reasonable amount of time, though further work will be needed to scale Replayer to larger formulas. We have demonstrated the validity of our approach and of our implementation prototype. While further work is necessary for our approach to scale, we believe that our initial results are promising. We discuss some possible strategies to improve the scalability of our approach in Section 6.6.

2006

"... In PAGE 8: ...744 15.29 Table4 : Checksum replay performance. For example, suppose that the original input io is a message to an SMTP server that causes the SMTP server to send an email message.... ..."

Cited by 16

### Table 4: Comparison of diagnostic approaches (293)

2001

"... In PAGE 30: ... QBC and Acridine orange and PCR are less promising for widescale use. Table4 compares the approaches. Despite numerous attempts it has not been possible to develop highly sensitive and specific clinical algorithms (125, 126), so that accurate diagnosis requires detection of parasites by microscopy or with immunochromatographic rapid diagnostic tests (RDTs) detecting parasite antigen.... ..."

### Table 5 Rafii and Kampas: Six Stages Framework [7] Stage Question

2005

"... In PAGE 9: ... For our analysis, we use a promising approach proposed by Rafii and Kampas [7]. They suggest that a disruption innovation process consists of six stages (see Table5 ). For each stage, there are contributing factors which enable or disable the disruption.... ..."

Cited by 3