"... In PAGE 4: ...METHODOLOGY The first research step consisted in drawing up a questionnaire composed of three parts. CIO competencies are evaluated through eleven sets of questions according to the framework shown in Table2 . Business performance (the dependent variables) is assessed through five sets of questions, corresponding to measures shown in Table 1: more precisely, this part of the questionnaire evaluates the extent to which the IS supports the improvement of each performance measure.... ..."

"... In PAGE 5: ... This trivial strategy gives zero recall to the positive class, and so achieves zero F-measure. Given the statistical machinery described in the previous section, the definition of the null hypothesis and the description of how to score each of the four performance metrics, we give in Table2 an explicit procedure for determining the statistical significance that the winner of a competition of C binary classifiers achieves score m on a test set comprising P positives and N negatives. The loop determines the required inverse CDF value empirically.... In PAGE 5: ... The loop determines the required inverse CDF value empirically. We note that the empirical method presented in Table2 is general for any performance metric and can be used without a need for analytical knowledge of the statistics behind a particular metric. Figure 1 shows the entire CDF for three of the performance metrics over a test set with P=10 positives and N=1000 negatives.... In PAGE 6: ...umbers, e.g. 380MB for C=1000. Generating the critical value charts in the following section consumed more than a year of CPU time, run on over a hundred CPUs provided by the HP Labs Utility Data Center. Calling the procedure outlined in Table2 for each performance metric and test condition would have required nearly 12 years of CPU time. To make this reference work feasible, a more efficient computation was performed, which computes the critical value for all performance measures and all values of C during a single run of R repetitions.... In PAGE 19: ...Table2 : Best Accuracy.... ..."

"... In PAGE 15: ... Table1 : Component and architectural competencies in high technology SME COMPETENCIES COMPONENT ARCHITECTURAL ASSETS Technological devices and dedicated equipment to research activity - KNOWLEDGE Specific knowledge Codified knowledge in specific disciplines Tacit knowledge in running experiments Diversified disciplinary knowledge General and abstract knowledge in several disciplines Tacit functional knowledge Tacit knowledge in project management (including networking) INDIVIDUALS Researchers Engineers Star researchers Managers MEASUREMENTS Patents and publication in a specific scientific area Circulation of information inside and outside the firm Table 2 : Component and architectural competencies in low technology SME COMPETENCIES COMPONENT ARCHITECTURAL ASSETS Machinery and equipment dedicated to production process - KNOWLEDGE Specific knowledge Technological knowledge Tacit knowledge in running production (including production failures) Diversified functional knowledge Knowledge about other technological devices that prove potential applicability (relationship with equipment suppliers and laboratories developing technologies) INDIVIDUALS Engineers Technicians Engineers Managers MEASUREMENTS Patents in a specific technological area Scientific advances in market Combination of existing technologies to other purposes; Ability of the manager to improuve production process Three important points are worth noticing : In low tech sectors as well as high tech, both component and architectural are needed to innovate. The nature of component and architectural competencies and the degree of its mobilisation are different.... ..."

"... In PAGE 36: ... This is called the ranking competence of the system, and demonstrates howwell the system makes appropriate distinctions between categories. The results of this comparison are summarized in Table2 . The rows represent the system interpretation that resulted in the highest association measure.... In PAGE 36: ... Figures 25 and 26 show a representative set of those shapes on which agreement occurred over the rst ranked category. Some of the disagreements shown in Table2 are primarily due to a person apos;s di culty in con- sidering a shape purely on function-based grounds. Representative examples of these types of discrepancies are displayed in Figure 28.... ..."

"... In PAGE 12: ... For example, if a case is removed from a NUN case-base then it will be the last case that was added. Results: The results are shown in Table2 . Each value is the mean competence of the case-bases produced by each of the editing strategies once they have been normalised to the appropriate RC case-base size.... In PAGE 12: ...ection 4.2 when consistency was measured. Moreover, the relative coverage measure performs well in both classification and CBR settings, while the NUN method performs relatively poorly in the CBR data-sets. In fact, in Table2 we see that the normalised competence values for the NUN case-bases are smaller than the competence values for the CNN case-bases, for the CBR data-sets.... ..."

"... In PAGE 2: ... He integrates these with scope and duration to define three distinct scales (cf. Table1 ): An emergency is a short-lived event whose effects are localized within a single community. The community as a whole and its response infrastructure remain fully functional, and its internal capacity is sufficient to manage the response.... In PAGE 4: ... Not surprisingly, partnership formation is essential to responses to events of all sizes except local emergencies, and grows increasingly critical as the scale of event increases (Quarantelli, 2005). Table1 summarizes measures of scale, and characteristics of events of different scales. Although differently-scaled events are qualitatively distinct, the scale of a particular event may not be apparent until the response is well underway (or even after it is concluded).... ..."

"... In PAGE 5: ...The procedure COMPETants, described in Table1 initializes the two ant colonies, deter- mines the population sizes, the number of spies for each population and controls the termination. In detail, for a number of max Iterations the procedure ACO is executed for the two ant colonies.... ..."

"... In PAGE 4: ... 4.4 Summary of Competencies Table3 summarizes the competencies identi#0Ced most fre- quently from the multiple sources. The Derived category refers to those competencies extracted from the analysis of the interview transcripts.... In PAGE 4: ... A number of competencies were identi#0Ced by the subjects and#2For managers, but were not included in the set of competencies that will be used for further research. Some of these rejected competencies overlapped with those in Table3 . However, most were rejected because few peo- ple identi#0Ced the competency,oritwas not validated by multiple sources.... ..."