Trends in expert system development: A longitudinal content analysis of over thirty years of expert system case studies.

Diagnostic expert system applications continue to be the most popular.Synthetic type problem domains (such as design and planning) tend you yield higher impact systems.Rule-based knowledge representations tend you yield higher impact expert system applications.Unstructured interviews are not commonly used now for knowledge acquisition.In the past, structured and unstructured interviews tended to yield systems with equally high impact. Research in Expert Systems (ES) has been one of the longest-running, and most successful areas of ongoing research within the AI field. Since the 1980s, many case studies of ES applications have been published covering a wide range of functional areas and problem domains. These case studies contain an enormous amount of information about how ESs have been developed and how the tools, concepts, and applications have evolved since their inception. This research has painstakingly collected and analysed the content of 311 ES case studies dating from 1984 through 2016. A detailed content analysis was performed on this corpus in order to capture as many details as possible from each case. Further value was added to the study by using an impact scale to try and gauge the impact or success of the resulting application. With such a large sample size, the results are helpful in identifying how ES research has evolved and areas for further research.