Abstract | ||
---|---|---|
This paper explores intelligent diagnostic support systems as debugging tools for end-users of computing. By analyzing error (fault) behavior of users and fault-diagnostic relationships of these errors, the authors identified patterns that could be exploited to provide electronic diagnostic assistance. This analysis showed that (a) error behavior differs considerably across end-users; and (b) individual end-users tend to make the same errors over time because they have difficulty identifying the causes of their errors. When viewed in light of the literatures on human-computer interface design and human error/diagnostic behavior, this analysis lead to some general conclusions about how diagnostic systems could be designed to provide better advice. Specifically, the empirical results suggest that diagnostic systems with firing rules based solely upon the aggregated behavior of all users will often provide individual users with poor advice. In contrast, diagnostic support systems could be improved by using user-specific data in the knowledge base. Such a deviation from conventional ideas about knowledge-base development seems consistent with other diagnostic situations, such as medical and machine diagnosis. |
Year | DOI | Venue |
---|---|---|
2004 | 10.1142/S0219622004001148 | INTERNATIONAL JOURNAL OF INFORMATION TECHNOLOGY & DECISION MAKING |
Keywords | Field | DocType |
decision support systems, expert systems, end-users of computing | Data mining,Diagnostic system,Support system,Computer science,Decision support system,Expert system,Human error,Artificial intelligence,Knowledge base,Machine learning,Debugging,Interface design | Journal |
Volume | Issue | ISSN |
3 | 3 | 0219-6220 |
Citations | PageRank | References |
3 | 0.40 | 14 |
Authors | ||
2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Vicki L. Sauter | 1 | 45 | 6.79 |
Laurence A. Madeo | 2 | 5 | 1.14 |