A Study on Rule Based Reporting Systems

Abstract

Government enacts rules & regulations. People are bound to follow those rules and regulations in their everyday life. Human brains do the computation for generating conclusions to take decisions on different circumstances applying those rules. A Human being does not need any external rule reasoner. However, nowadays data is generated by different automated systems, and inserted, updated, deleted in different database design formats. So, the decision-making becomes difficult by generating required conclusions from large repositories of data having multiple formats. Therefore, artificial intelligence in form of the Rule Based System can help to mitigate this problem by analyzing the conclusions from the given facts. But, the existing Rule Based Systems processes the adverse events from only one data sources at a single time and manually generates the facts. In this circumstance, the study proposes an automated Rule Based System, which is capable of generating conclusions as output for the given input queries. The study results in a database independent Rule Based method with dynamic predicate generation & translation for analyzing the stored adverse events from multiple databases by applying the predefined set of rules and regulations. The formal rules are generated using Defeasible Logic (DL) to efficiently handle the logic-based implementation of the proposed methodology. In addition, the proposed system includes SPINdle rule engine to generate required conclusions by loading the formal rules that are generated using the DL. The resulting Rule Based Reporting System (RuleRS) integrates simultaneous reasoning generation from the adverse events along with the existing Rule Based Systems to get the argumentation from multiple relational databases. It also supports the easy integration of additional data sources for generating assumptions. The output is sent to the I/O section for showing the generated reports. The study also conducted an empirical evaluation for the single sources (i.e., FAERS & ChildSafe database separately) and proposed RuleRS in reasoning simultaneously from multiple sources of data (i.e., FAERS & ChildSafe database concurrently). The evaluation result is promising to integrate the developed system with the existing RuleRS.