A Comparison between Similarity Metricies in Case-Based Reasoning Systems

Abstract

Mobile bank applications provide many services to customers and this study supports the model of simulation of the application's performance. There are many problems threatening applications, such as reliability, connection and confidentiality problems, and problems related to the system’s behavior during service interruptions or lack of availability at the time specified for the customer, and these are the main problems that affect the reliability of the system. In This research, the main concept is autonomic system with its different parts in general ,and self-healing in more details, It Provides a mechanism for self-Healing using an inference algorithm based on previous cases (Case-based reasoning). The algorithm enables the model of simulation system to detect and identify the problem and then self-treat it, and it is able to examine and treat itself without external interference. This mechanism is applied to the mobile bank application as a case study of an application that contains sensitive data. This algorithm was implemented using Eclipse using Java language and used the case database to save a set of cases similar to the occurrence of the problem, The algorithm selects the best solution to solve the problem among the set of available solutions using a set of functions to calculate the similarity distance between these cases and accuracy and error rate.