Fortifying Abridgement Assessment System Against Student Collusions by Combining Fast Text Similarity Computation and Disjoint Set Union

Various forms of assessment of reading comprehension of students have been and are being practiced worldwide, but limitations apply when it comes to pursuing quantifiability and computability. Abridgement method, however, is undergoing rapid developments under the recent researches, demonstrating potentials to resolve these drawbacks. This method has allowed for handy implementation with plenty of extra features that assist educators throughout the assessment procedures, letting them play a more important role in the field of education. Nevertheless, the system is found partially hackable at this early stage of development despite its usefulness, potentially leading to a failure to detect academic misconduct such as collusion between students while marking. This research, in response to raised concerns about those issues, proposes the invitation of a data structure in graph theory called disjoint set union, accompanied by the deployment of external library that computes text similarity at very high speed, to the system. This additional layer of the assessment architecture is expected to enhance its rigidity, contributing to the sophistication of this educational method itself.