National Examination (UAN or UNAS or UN) is a government evaluation measurement tool used to determine the quality of education in Indonesia. Quality is evidenced when students can answer questions in the national examination based on materials aligned with the Graduate Competency Standards (SKL). Materials published in the past can also be republished for the next exam. This SKL is divided into several research topics that are relevant to current trends. This study aims to determine how effective the use of the Generalized Vector Space Model (GVSM) algorithm is in predicting each question that will appear. The GVSM algorithm determines the similarity between words that appear in one document and words that appear in another. The GVSM algorithm is used to determine the similarity between words that appear in one document and words that appear in another document. The evaluation results showed an accuracy of 0.75, a precision of 0.7321, and a recall of 0.7017. © WCSE 2025. All rights reserved.