Results 11 - 20 of 5,737

Table 1. Comparison of the percent error for backprop (BP) and for BCM/BP with the isolated stressed vowel task. Stops We used two di erent stop consonant token sets, both with 16ms resolution. The rst set used only stop tokens that were sequences of stop-closure followed by stop-release (CLREL tokens). The second token set included all in- stances of stop-release, reguardless of what the prior con- text was (REL tokens). Tokens from the rst set are more

in Phonetic Classification Of Timit Segments Preprocessed With Lyon's Cochlear Model Using A Supervised/Unsupervised Hybrid Neural Network
by Gary Tajchman , Nathan Intrator

Table 15.2 summarizes the best frequency transformations for the syntagmatic uses of context. As with the context regions, the best transformations are more or less consistent across the evaluation metrics. The exceptions are small context regions for the thesaurus comparison, large regions for the association test, and large regions for the part-of-speech test. For the thesaurus test with small con- text regions, binary counts produce the best result for the strict evaluation, but dampened frequency transformation produce the best result for the lax one; for the association test with large context regions, dampened frequency transforma- tion produce the best result for the strict evaluation, but tfidf-transformation produce the best score for the lax setting; for the part-of-speech test with large context regions, dampened frequency transformation produce the best results for the strict evaluation, while binary counts lead to better results for the lax setting.

in The Word-Space Model Using
by Magnus Sahlgren, Isbn Nr 2006

Table 1. Example of context presented as boolean matrix; two-class (+ and -) data indicated by class column.

in CLANN: Concept Lattice-based Artificial Neural Network for supervised classification
by Norbert Tsopzé, Engelbert Mephu Nguifo, Gilbert Tindo, Yaoundé Cameroun
"... In PAGE 3: ... transactions). Table1 is an example of con- text. We denote in the next sections the set fa; b; c; dg (resp: f1; 2; 3; 4g) as abcd (resp.... ..."

Table 5: Estimated CAP Context # of CAPs

in Example-based Machine Translation without Saying Inferable Predicate
by Eiji Aramakiýþ, Sadao Kurohashiýþ, Hideki Kashiokaþ, Hideki Tanakaýýý
"... In PAGE 6: ... It shows that we can extract one CAP from each 12 sentence pairs. Table5 shows the ratio of their estimated con- texts. Most of them were estimated as BOTH- CONTEXT.... ..."

Table 1 Analytic Centers

in The education testing problem revisited
by Moody T. Chu, Joel W. Wright
"... In PAGE 12: ... The iteration stops when the 2-norm of the di erence between consecutive iterates is less than kSk where is the machine dependent oating point relative accuracy and is 2:2204 10 ;16 in our case. In Table1 we list the column indices used to generate the matrix S, the analytic centers and the number of iterations needed for convergence. So as to t the data comfortably in the running text, we display all the numbers with only four decimal digits.... ..."
Cited by 1

Table 1 Analytic Centers

in unknown title
by unknown authors
"... In PAGE 12: ... The iteration stops when the 2-norm of the di erence between consecutive iterates is less than kSk where is the machine dependent oating point relative accuracy and is 2:2204 10?16 in our case. In Table1 we list the column indices used to generate the matrix S, the analytic centers and the number of iterations needed for convergence. So as to t the data comfortably in the running text, we display all the numbers with only four decimal digits.... ..."

Table 1. Order 3 context for the example in Figure 1, constructed by means of the inductive algorithm.

in Object Oriented Design Pattern Inference
by Paolo Tonella, Giulio Antoniol 1999
"... In PAGE 5: ... Let us consider the example in Figure 1. If classes B; C; Y and Z own method f, the associated con- text has to be augmented with the three attributes ohfi(1); ohfi(2); ohfi(3), since the classes owning f may occupy all the three available sequence positions in the con- text given in Table1 . By applying concept analysis to the resulting context, three design patterns are obtained after simplification.... ..."
Cited by 20

Table 1. Order 3 context for the example in Figure 1, constructed by means of the inductive algorithm.

in Object Oriented Design Pattern Inference
by Paolo Tonella, Giulio Antoniol 1999
"... In PAGE 5: ... Let us consider the example in Figure 1. If classes B; C; Y and Z own method f, the associated con- text has to be augmented with the three attributes ohfi(1); ohfi(2); ohfi(3), since the classes owning f may occupy all the three available sequence positions in the con- text given in Table1 . By applying concept analysis to the resulting context, three design patterns are obtained after simplification.... ..."
Cited by 20

TABLE 2. Analytical expressions for the three models discussed in the text.

in Observations of Vela X-1 with RXTE
by P. Kretschmar, I. Kreykenbohm, R. Staubert, J. Wilms, M. Maisack, E. Kendziorra, W. Heindl, D. Gruber, R. Rothschild, J. E. Grove

Table 4 Comparison of the results of all the algorithms for the two test texts.

in Natural language tagging with parallel genetic algorithms
by Enrique Alba, Gabriel Luque
"... In PAGE 7: ... In general, the results obtained for this algorithm are very poor, indicating that SA is not able to solve this problem adequately. Table4 presents the best value and the average for the conflguration which provides the best re- sults of each algorithm, i.e.... In PAGE 7: ... First we compare our algorithms, and latter, we compare our results with the Viterbi ones. We can observe in Table4 that the GA has reached the globally best results for all the test texts and con- texts, though the difierences are small. This proves that the exploration of the search space given by the classical crossover and mutation operators are enough for this speciflc problem.... ..."
Cited by 1
Results 11 - 20 of 5,737