A rule-based classification of short message service type

Short message service (SMS) is one of the most popular means of communication due to its type of usages like one-time passwords(OTPs), banking transaction alerts, and other promotional messages. SMS is the most time-sensitive channel of communication that demands service providers to send any information globally without any delay in respective time zone. For successful implementation of this service, the Telecom Regulatory of India (TRAI) has recommended certain guidelines that must be followed for the delivery of short messages. The prevalent practice in the industry is to store rules or raw string in a database and to match with incoming SMSs in synchronous mode. However, there are various shortcomings in this traditional approach that have been also discussed in this paper. Further, to address these issues, this research work proposes a novel approach for automated classification of SMSs in real time by the SMS service using a rule-based system of database template matching. Further to validate the proposed approach, SMS database of Netcore Solutions Pvt Limited, India, has been used for training the proposed algorithm. Proposed rule-based classifier keeps learning and updating the training dataset as more SMSs are accumulated and processed through the server. Other advantages, in terms of performance metrics, of the classification using the proposed rule-based database template matching over the traditional database matching approach have been reported in this work as evaluation measures. Reported rule-based SMS classification algorithm shows highest average classification accuracy of 100% which is better as compared to the similar research works already available in the literature.