Action rules are based on congruent predicaments and strategies that can be triggered when possible transitions of objects occur from one stage to another. This study presents a new way of implementing actionable discoveries in regards to object-driven approaches. To manifest usefulness of this new strategy in clinics, experiments were carried out in a new algorithmic database system in support of diagnoses of patients with liver disorders. The presented algorithm utilizes and expands the realization of action rules in which medical decision processes can be properly made for a clinical decision support system. The main features of the approach are: i) to utilize the discovered action rules, which are parallel to a bottom-up strategy that formulates the rules with a condition in minimal length; ii) to generate an object-driven action rule mining by the contingent algorithm where the data is processed by the expert system, producing actionable patterns due to object-driven action rules; iii) to discretize the data of the selected patients and extract the highest related attributes in test values; and iv) to validate the results along with the patients’ history and physical examination. Object-driven approach is a shortcut of the DEAR algorithm in which classification rules produced design patterns and therefore their results are limited. On the other hand, the method of object-driven of expert system, where the rules are combined and actionable patterns are shifted through in terms of breath-first traversal and redundancy is minimized. As a result, the object-driven approach is more robust and faster that means it reduces the imputing time and tautology.
Prof. Dr. Bilal BİLGİN