A Beacon based Monitoring system was developed to track whether the person or elderly who stay at home or not. However, there were some issues with false alarms from the alarm management system. Alarms need to be analyzed to determine patterns and trends of false alarms to reduce the number of false alarms in the future. In addition, rules for determining when an alarm is triggered are difficult to fine tune without detailed data analysis. This results in parameters that are set too wide, allowing for unnecessary alarms and cases to be created, which will need to be checked and verified, causing resources to be allocated for redundant cases. Therefore, this seed project aims to understand and evaluate patterns and trends to reduce the number of false alarms by an IoT deployment and extrapolate to filtered out with similar connectivity patterns. By understanding how connectivity patterns are related, the reduction of false alarms will reduce the resources needed to follow on verification. Such an approach would be helpful to the welfare’s workers who need to take care the elderly who stay home alone as the reduction of false alarms reduce their workload significantly. To achieve this, a false alarm detector will be developed based on the analysis of past data from the existing system. Using machine learning and AI, the false alarm detector will derive patterns between alarms generated in the two data sets to develop criteria for false alarms. Using the simulation tool and the AI algorithm, testing can be conducted according to future situations to reduce the number of false alarms.