The INO-ICAL prototype stacks have been extensively used to establish performance benchmarks of Resistive Plate Chambers for the final ICAL detector. Various studies have been undertaken using the data recorded by these prototype detectors. However, the studies primarily focused on single muon events. The existing traditional algorithms used for event analysis discard any events that exhibit a different topology from a straight line, assuming such events to be noise. While this approach leads to the loss of many interesting events in the data, it can also reduce the physics potential of the detector. In this context, we have developed a ML model to identify the new event topologies, specifically those identical to muon electron elastic scattering events. We present the overview of the algorithm and the potential outcomes from this study.