Tools
Sorted by:
Try your query at:
 |
 |
 |
 |
 |
 |
Results 1 - 10
of
32,170
Table 2: Resultant Correlation Table for the KWC Example During instantiation of correlation rules, uncertainties in the correlation table (e.g., +{, { { ) should be resolved. For example, the software architect could consult domain expert who knows about the characteristics of the intended application domain in order to determine if the Pipe and Filter would signi cantly hinder the Space Performance goal. Once obtained, domain characteristics can be used as an argument (e.g., support4: expected size of data is huge (from domain expert), as shown in Figure 5).
1995
"... In PAGE 8: ... The software architect has instantiated correlation rules to establish both positive and negative links, examined them, rejected or tailored some of them, and provided justi cations. Based on Table 1 described earlier in Section 2, Table2 illustrates a resultant correlation table, as an alternative representation to what is shown in Figure 5, which might aid visual understanding when the number of correlation links is high. Performance Deletability [Function] Modifiability[Function] Shared Data Implicit Invocation Pipe amp; Filter Comprehensibility Modifiability [Process] Simplicity Coherence support2: [Parnas72] support3: fewer asumptions among interacting modules support2 support4 deny1 [System] Modifiability [Function] Extensibility [Function] Updatability Reusability [Data Rep] support1 PerformanceTime ! ! ! ! ! ! ! Space Performance ! ! ! ! ! Modifiability ? ? ? ? ? support1: among the vital few goals (from market survey) support4: expected size of data is huge (from domain expert) deny1: many implementors familiar with ADTs (from domain expert) Legend weak negative satisficing strong negative satisficing Evaluation Labels Link Types Criticality + ++ + -- -- + ++ + + ++ -- -- + + + + + + + + + support3 -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- ? ? ? -- Abstract Data Type ++ + weak positive satisficing strong positive satisficing very critical ! ! ! critical denied neutral undetermined satisficed ? Figure 5: Goal graph selecting among architectural alternatives for a KWIC system.... ..."
Cited by 18
Table 1. Expert system applications
"... In PAGE 6: ... The emphasis was strongly on practical applications; very little basic AI research was done. Table1 lists the most influential and widely used ones of these applications, some of which have been in use for close to 15 years. The table also shows the domain, the lifetime, the software and hardware environment (which may have changed during the lifetime), and the current status of each application.... ..."
Table 4. The evaluation of the relevance and objectivity of knowledge sources by a transaction domain expert
"... In PAGE 6: ... A selected set of knowledge sources for the overall solution domain KS KNOWLEDGE SOURCE KS1 Concurrency Control amp; Recovery in Database Systems [17] KS2 Atomic Transactions [21] KS3 An Introduction to Database Systems [18] KS4 Database Transaction Models for Advanced Applications [19] KS5 The design and implementation of a distributed transaction system based on atomic data types [25] KS6 Concurrency Control Performance Modeling: Alternatives and Implications [13] KS7 Principles of Transaction Processing [16] KS8 Course Notes of Transaction Design KS9 Concurrency Control in Advanced Database Applications [15] KS10 Conference Proceedings on Advanced Transaction Systems and Applications KS11 On-Line Transactions Tutorial [24] KS12 Transaction Domain Expert with 15 years of experience KS13 Design of Adaptable Transaction Systems [23] KS14 Nested Transactions [22] KS15 Adaptable Concurrency Control for Atomic Data Types [14] KS16 A survey of techniques for Synchronization and Recovery in Decentralized Computer Systems [20] 5.2 Evaluation by a Domain Expert Table4 represents the evaluation of relevance and the objectivity of the knowledge sources by a transaction do- main expert who has experience in the theory, design and implementation of a wide range of transaction sys-... ..."
Table 1. Distribution of domain and methodological experts by site.
"... In PAGE 14: ... Results We performed analyses on 493 projects in the sample, omiting one outlier project from the sample of 494 because it was very large, had extremely high net earnings, and biased the data to favor our hypotheses. Organization of expertise acros sites Table1 shows the distribution of expertise by sites, indicating that domain experts dominated methodological experts at each site. This imbalance reflected customer requirements.... In PAGE 14: ...ethodological experts at each site. This imbalance reflected customer requirements. Of the 493 projects, 279 projects (57% of the projects) required domain expertise only, 17 (36%) required both domain and methodological expertise, 24 (5%) required methodological expertise only, and 13 (3%) required uncoded expertise, such as management of art colections. Table1 about here Figure 1 shows how domain and methodological expertise was distributed acros all six sites in year 200. Each circle represents a site, its size approximately proportionate to the number of professionals.... ..."
Table 1 III.2.2 Implementation and Experimentation Due to the availability of a domain theory in the form of expert rules, strategies I and II seemed best adapted to the toxic coma application. The first step consisted in testing the efficacy of knowledge- based configuration. Prior knowledge came in the form of three rule sets concerning tricyclic antidepressants (A), phenothiazines (P) and carbamates (C). In each case, expert rules concluded only on the presence in isolated form of the target toxin; there were no rules for detecting the presence of these toxins in combination with others, nor for concluding the absence of the target toxin. Each rule set was mapped onto a neural network, which was then trained on the given data using ten-fold stratified cross-validation. The mean accuracy rate measured on the test sets was then compared to the baseline accuracy, i.e., the accuracy attained by the default strategy of simply assigning each
"... In PAGE 30: ...2.2, both the prior knowledge and the available data for the toxic coma problem were deficient ; thus the NeuroTox system was naturally led to adopt Strategy IV shown in Table1 . Final results of the experiments described above are shown in Table 4.... ..."
Table 1. An Intuitive Domain Model of Expert Finding Systems Basis for
2002
"... In PAGE 9: ... A domain factor together with its actual value(s) is called domain characteristic. For the expert finding systems domain, we identified seven facets/domain factors as shown in Table1 . Each column represents one facet/domain factor and... ..."
Cited by 12
Table 1. An Intuitive Domain Model of Expert Finding Systems Basis for
"... In PAGE 9: ... A domain factor together with its actual value(s) is called domain characteristic. For the expert finding systems domain, we identified seven facets/domain factors as shown in Table1 . Each column represents one facet/domain factor and... ..."
Table 1. An Intuitive Domain Model of Expert Finding Systems
2002
"... In PAGE 17: ... A domain factor together with its actual value(s) is called domain characteristic. For the expert finding systems domain, we identified seven facets/domain factors as shown in Table1 . Each column represents one facet/domain factor and its possible values, which are possible implementations of the respective domain factor.... ..."
Cited by 12
Table 1: Security assessment environments. Adapted from [6] Environment Time scales Typical problems Operator Expert
1997
"... In PAGE 5: ... 2.3 Practical application domains Table1 shows the practical study contexts or environments which may be distinguished in security assessment applications. The first column identifies the study context; the second specifies how long in advance (with respect to real-time) studies may be carried out; the third column indicates the type of subproblems that are generally considered in a given environment; the last two columns indicate respectively if an operator is involved in the decision making procedure and if an expert in the field of power system security is available.... In PAGE 6: ... Operation planning. As suggested in Table1 , operation planning concerns a broad range of problems, including maintenance scheduling (one year to one month ahead), design of operating strategies for usual and abnormal situations, and setting of protection delays and thresholds. The number of combinations of situations which must be considered for maintenance scheduling is also generally very large, and automatic learning approaches would equally be useful to make better use of the available information and to exploit the system more economically.... ..."
Cited by 4
Results 1 - 10
of
32,170
