Assessment of Subject Competitiveness Using Node Coverage in a Temporal Knowledge Graph

In this paper, we present a new method for analysing students' ability in algorithmic competitions and their potential for subsequent development. We create an innovative prediction method based on the properties of time-series knowledge graphs for predicting students' performance in subsequent algorithmic competitions. The method is based on constructing a time-series knowledge graph based on data from students' previous problems, which is then further analysed and predicted. To collect the data, we refer to a large number of mainstream authoritative competition platforms, which are then used to characterise each participant's performance in the problem solving process. The characterisation included the difficulty of the problem and the algorithms involved. The aggregation of mastered and unmastered algorithms, combined with the analysis of the obtained TKGs, helps to derive the current level of student competition. By aggregating the mastered and unmastered algorithms and analysing the obtained TKGs, we can derive the student's level of competitiveness and analyse and calculate the student's potential for subsequent development.