The GRAPES-3 experiment located in Ooty consists of an array of 400 plastic scintillator detectors spread over an area of 25000𝑚 2 and a large area (560 𝑚 2 ) muon telescope. Every day, the array records about 3 million showers induced by the interaction of primary cosmic rays in the atmosphere. One of the primary objectives of the experiment is to measure the energy spectrum and composition of the cosmic rays in the TeV-PeV energy range. However, some of the detected showers have cores outside the array. This fraction increases with energy due to the higher lateral spread of shower particles at higher energies. Identifying these events is thus crucial for accurate measurement of the cosmic ray energy spectrum. This work will describe simple cut based as well as machine learning based strategies for identifying and excluding such events and their impact
on the cosmic ray energy spectrum as measured by the Bayesian unfolding technique.