"... In PAGE 5: ... Therefore, the node refinement improve the computational throughput. In Table3 shown the global h-refinement and local h- refinement to reach the same l2 error. This corresponds to... In PAGE 5: ... Comparison between global h- refinement and local h refinement. In the second and fourth column of Table3... In PAGE 6: ...in the number of nodes to reach the same root mean square, in comparison with the scheme of global h-refinement. In order to meet the same l2-error, see Table3 , the local h-refinement scheme requires an 82% less nodes in com- parison with the required nodes of the global h-refinement technique. In the third column of Table 3 it is observed that the norm l1 is approximately one order of magnitude better than the obtained with the global refinement.... In PAGE 6: ... In order to meet the same l2-error, see Table 3, the local h-refinement scheme requires an 82% less nodes in com- parison with the required nodes of the global h-refinement technique. In the third column of Table3 it is observed that the norm l1 is approximately one order of magnitude better than the obtained with the global refinement. In order to achieve a similar error with the global refinement, we require of 2700 nodes distributed uniformly, obtaining an error l1=8.... ..."

"... In PAGE 28: ... For example, fORGANIC CHEMICAL, LIPIDg in the lexical metaschema is a refinement of ORGANIC CHEMICAL in the consensus metaschema. Table9 shows all such refinement cases. This kind of refinement covers 16 semantic types.... ..."

"... In PAGE 6: ... Then, after the refinement, the class chosen to compute the identification rate is the one, none discarded by the method, having the greatest probability. Table3 lists the results obtained after the refinement of the classification.... ..."

"... In PAGE 5: ... Preliminary classification results the three types of plots. Table2 lists the results after the refinement of the classification has been applied to the 5 images of the series.... ..."

"... In PAGE 28: ... Table8 shows the cases of refinement from the lexical metaschema to the consensus meta- schema. The size of a meta-semantic type is displayed in parentheses following the name.... ..."

"... In PAGE 10: ... Then, after the refinement, the class chosen to compute the identification rate is the one, none discarded by the method, having the maximum probability. Table1 explains how the refinement and decision tasks are processed on a specif- ic plot. The first line represents for the five images I1 I5 the classes given by the preliminary classification.... ..."

"... In PAGE 5: ... Then, after the refinement, the class chosen to compute the identification rate is the one, none discarded by the method, having the maximum probability. Table3 lists the results obtained after the refinement of the classifi- cation. The result clearly shows the progression of the number of clear plots on all images (up to 54:3% for image I3).... ..."

"... In PAGE 29: ... As an example, fPHENOMENON OR PROCESS(4), NATURAL PHENOMENON OR PROCESS(1), BIOLOGIC FUNCTION(1)g in the cohesive metaschema is a refinement of PHENOMENON OR PRO- CESS(6) in the lexical metaschema. Table6 shows these three cases of refinement of the lexical metaschema which cover 34 semantic types, in both metaschemas.... ..."

"... In PAGE 8: ... LOOP is currently under development. We summarize our comparisons in Table1 . In compari- son with other tools, PD offers a software oriented approach to refinement rather than a brutally mathematical one.... ..."

"... In PAGE 12: ... However, since the density is not perfectly clear, it has been modeled as solvent. The final phasing and refinement statistics are given in Table2 . Co- ordinates and structure factors have been deposited at the PDB with access code 2BI0.... ..."