"... In PAGE 51: ... One of these four fuzzy datasets is considered as an historical dataset, the other three are the current datasets containing the new projects. Table1 shows the results obtained using only the max-min aggregation to evaluate the individual distances (Formula (1.1)).... In PAGE 52: ...Table1... ..."

"... In PAGE 10: ...2.1 Distance Measures Table5 compares total current delivered across the eight different distance measures we employed. First, we consider distance measures for candidate gener- ation, show in Table 5.... In PAGE 10: ....2.1 Distance Measures Table 5 compares total current delivered across the eight different distance measures we employed. First, we consider distance measures for candidate gener- ation, show in Table5 . We observe that as the algorithm is given more resources (ie, as the stopping condition changes), the best distance measure also changes.... ..."

"... In PAGE 10: ...2.1 Distance Measures Table5 compares total current delivered across the eight different distance measures we employed. First, we consider distance measures for candidate gener- ation, show in Table 5.... In PAGE 10: ....2.1 Distance Measures Table 5 compares total current delivered across the eight different distance measures we employed. First, we consider distance measures for candidate gener- ation, show in Table5 . We observe that as the algorithm is given more resources (ie, as the stopping condition changes), the best distance measure also changes.... ..."

"... In PAGE 2: ... Currently, ITTACA focuses on three types of cancer, for which the data integration has been as exhaustive as possible: breast carcinoma, bladder car- cinoma and uveal melanoma. Table1 lists all of the public datasets currently available in ITTACA (12 article references: 2 for bladder carcinoma, 8 for breast carcinoma and 2 for uveal melanoma). Datasets are selected based on the availability of relevant anatomo-clinical data.... ..."

"... In PAGE 5: ... Our training and testing in this area was limited however, because our dataset did not contain many instances of the same attack. Table1 shows some of our results where sshprocesstable is the name of a particular type of denial of service attack. In Table 1, columns 2, 3, 4, and 5 respectively refer to the correct prediction of normal traffic intensity, the incorrect prediction of normal traffic ... In PAGE 5: ... Table 1 shows some of our results where sshprocesstable is the name of a particular type of denial of service attack. In Table1 , columns 2, 3, 4, and 5 respectively refer to the correct prediction of normal traffic intensity, the incorrect prediction of normal traffic ... ..."

"... In PAGE 19: ...um.pos/Num.neg records Dimensions positive Ideal 10000 2 5000 1 Diag2d(10%) 10000 2 5000 1 Diag2d 100000 2 50000 1 Diag3d 100000 3 50000 1 Diag10d 100000 10 50000 1 Noise2d 10000 2 5000 1 7.2 Real-world Datasets We used UCI amp; KDD data (listed in Table2 ), but we also experimented with datasets of particular current interest within our laboratory. Life Sciences.... ..."

"... In PAGE 12: ... In our experiments, we use the same 26 datasets from UCI Machine Learning Repository [Merz96] as in [Liu98]. The detailed information about these datasets is listed in Table1 . Since the current version of our algorithm can only deal with nominal attribute, the entropy-based discretization algorithm [Fayyad93] is used for preprocessing.... ..."

"... In PAGE 24: ... 24 expert has 25 neurons in the hidden layer. In Table1 there are the confusion matrices obtained on the TS by the best audio, camera motion and face experts. The face expert required a different experimentation since it employs a naive classifier.... In PAGE 26: ... 20.01% 79.99% 23.59% 76.41% 30.45%69.55% Table1 . The confusion matrix obtained on the TS by the best audio, camera motion and face expert, where ND and D stand for Not-dialogue and Dialogue shot, respectively.... In PAGE 28: ... It was also an objective to be able to study the evolution over different time periods, so identification of the age/date of the publication/collaboration link was collected. 4 Table1 has the number of publications or scholarly works that the researchers of the various departments have been a part of. This total of 1781 is different from 1480 because authors A and B can be from two faculties and have worked on the same publication P.... In PAGE 28: ... 4. FINDINGS Table1 summarizes the data by faculty of the total number of publications per researcher, the number of unique coauthors and the total number of coauthor collaborations. In the dataset the average number of publications per researchers is 19.... In PAGE 28: ...21. One objective of this project was to find out whether the data summarized in Table1 would in more detail characterize the researcher network at MC as a small world network. In summary we find it can be called a scale free network with some characteristics of a small world network.... In PAGE 29: ... 4 Betweeness ranking vs. #Unique Co-Authors in Figure 1 and summarized in Table1 , the standard deviation from the average number of unique coauthors is broad. Another characterization of networks is that they change or evolve over time.... In PAGE 32: ... [14] We can look at the association of the publication count to past collaborations. Table 5 shows that while all departments are collaborating more often externally, with coauthors in an outward direction rather than within the school, the Logistics department has the greatest ratio and the greatest number of publications in Table1 . This would indicate that external collaboration helps in the production count.... In PAGE 57: ... 57 Table1 : Experimental Conditions (installed or not installed) Case Node1 Node2 Node3 Node4 Node5 No.0 No No No No No No.... In PAGE 57: ...2 No Yes Yes Yes Yes No.3 Yes Yes No Yes Yes The response time is measured for four conditions, shown in Table1 . Nodes 1-5 are the computers, and cases 1-3 are the installation conditions.... In PAGE 60: ... BACKGROUND OF PROBLEM 3.1 Higher Education System in Taiwan Under national educational policy, there are presently two types of higher education institution: academic universities and technological institutions ( Table1 ). Academic universities provide a broad range of fields of study including arts, agriculture, architecture, commerce, engineering, law, mathematics, medicine, natural sciences, social sciences, and veterinary science that are similar to those provided by universities in the U.... In PAGE 61: ... The advantages of e-learning are clear in the area of timely delivery of training materials, increased convenience, and increased learning effectiveness [28], [29]. English is the de facto language of the Internet as of September 2003 ( Table1 0). In other words, web sites written in English are in a large majority on the Internet.... In PAGE 61: ...biz/globstats/index.php3 Table1 0: Online Languages All students in Taiwan must take English courses from seventh grade to twelfth grade at least. For convenience of learning, broadening of educational resources, and development of students with an understanding of international issues, to use the Internet, to conduct Internet searches, and to read online journals and documents in English is the optimum approach to achieve these goals for Taiwanese students.... ..."