"... In PAGE 6: ...3 NLP Software Stemmers, morphological analyzers and MT systems have been widely used by the participants. The list of tools can be seen in Table3 . Some results are worth pointing out: { The best groups in the German monolingual retrieval task all did some kind of compound analysis, con rming that morphological information (beyond crude stemming) may be crucial for languages with a rich morphology.... ..."

"... In PAGE 38: ...ocuments i.e., 1,980 reviews were examined (Experiment A). As shown in Table5 , the average accuracy of the NB text categorization on the original 11 classes is 19.55%.... In PAGE 39: ...s. Stimulating 1 vs. Stimulating 2. Descriptive statistics and the results of these experiments are reported in Table5 . The superclass classification experiments ran at rates much better than the baseline, providing evidence that the superclasses identified in the aforementioned clustering experiment are not arbitrary and the unigrams in our review data can predict suggested usages at the superclass level.... ..."

"... In PAGE 8: ... The compiler uses a flow- sensitive intra-procedural pointer analysis, which affects the precision of the disambiguation. Table1 shows how often scalar promotion can be applied. It separates promotion according to address kind (loop-invariant versus strided) and according to the control-flow constraints (o, from old , indicating whether the code could be handled by the traditional algo-... ..."

"... In PAGE 6: ... We selected these ve protocols because they covered a wide spectrum of options available to system designers. Table2 summarizes the design parameters of the four protocols. The rest of this section describes the protocols that we evaluated, including the message count that was used to simulate various consistency operations.... ..."

"... In PAGE 3: ... Such a distance matrix can be used as input to a clustering algo- rithm. The distance matrix for Table1 is shown in Table 2. Note that any relation the variables had with the programs... In PAGE 6: ...s the concept of those items having feature P1, i.e., the fields used in program P1. All concepts that can be identified from Table1 are summarized in Table 3. The items of a concept are called its extent, and the features its intent.... ..."

"... In PAGE 8: ....1. Applying concept analysis The case study started with the analysis of the source code to identify the files of the system and the programs using them. As result of this analysis, a file usage table containing 63 rows (programs) and 54 columns (files) was built (see Table1 ). Concept analysis was applied to this table and produced the concept lattice shown in Fig.... ..."

"... In PAGE 3: ....1. Core Promoter A core (or basal) promoter is located between approximately -40 and +35 relative to the TSS of the metazoan genes [4]. Four major core promoter elements have been identified: a TATA box, an initiator element (Inr), a downstream promoter element (DPE), and a TFIIB recognition element (BRE) [4, 8] ( Table1 ). The BRE is bound by TFIIB while the other three elements are bound by components in the TFIID complex (Table 1).... In PAGE 3: ... Four major core promoter elements have been identified: a TATA box, an initiator element (Inr), a downstream promoter element (DPE), and a TFIIB recognition element (BRE) [4, 8] (Table 1). The BRE is bound by TFIIB while the other three elements are bound by components in the TFIID complex ( Table1 ). The TATA box (also called a Goldberg-Hogness box) was the first core promoter element identified in eukaryotic protein-coding genes [10].... In PAGE 3: ... Comparison of promoter sequences from transcribed protein-coding genes revealed that most of them contain an adenosine (A) at TSS (+1) and a few pyrimidines surround this nucleotide [11]. This 7-8 bp motif ( Table1 ) was defined as a discrete core promoter element and named as the initiator. The initiator functions similarly, with regard to transcription initiation, to the TATA box and often independently of a TATA box.... ..."

"... In PAGE 3: ... The DNA binding motifs were retrieved from the Trans- Fac database. For both target genes, we found DNA binding sites for the predicted modulator in the 1000 bp upstream sequence ( Table2 ). The core similarity scores (CSS) and matrix similarity scores (MSS) that measure the quality of the match are all equal or very close to 1.... ..."

"... In PAGE 7: ... It is apparently better to divide the memory into several paths as described above, than using all memory for one single path with a large table. Table3 shows different configurations, all with a total memory of 4M entries, when 512k identifiers are in- serted, and the number of paths are varied. By adding more paths the overflow is effectively reduced.... ..."

"... In PAGE 11: ... The usage of the service has been observed for one year. Results of the usage analysis are summarized in Table2 . In particular, the first column lists the character- istics which were analyzed.... In PAGE 11: ...Identification of who and when used anonymous Web users from anonymous Web users from community-driven matching service more than 25 countries more than 40 countries Table2 demonstrates (as expected) a relatively infrequent usage of the system just after its launch. For example, during the first half of the exploitation period no new (to the system) ontology namespaces were acquired, namely all 8 namespaces already existed in the ontologies offered to the prototype users as examples.... ..."