Detection algorithm of large-scale collusion in M.C.Q. examinations

Collusion during M.C.Q. examinations is one of the vices faced by every educational system consistently over the years, especially during the COVID-19 pandemic, with the growing prevalence of non-proctored online examinations. With the advancement of technology and the usage of mobile phones, students located at far distances can also communicate during examinations. The proposed method provides an advancement to the already existing means to identify the groups of potential cheaters at any fixed level of confidence without the necessity of seating arrangement based information. This paper also gives a novel and robust quantitative measure to determine the extent of interaction within the suspected groups within which cheating has occurred. The procedure in this paper relies on a newly defined Gamma Index which provides a new framework for quantification of the similarity between answers of a pair of candidates