Trend analysis in competitive programming using topic modeling

Abstract

Competitive programming (CP) is a mind sports activity where people solve problems using command-line computer programs to provide correct output for the given test cases. Competitors need to practice problem-solving and mathematics as well as study algorithms and data structures to perform well in CP. This study aims to provide an original way to perform a trend analysis in CP, distinguishing topics frequently used in CP contests. To fulfill our goal, we create topic models based on previous topic modeling works to do natural language processing tasks using Latent Dirichlet Allocation (LDA) and Biterm Topic Model (BTM). For our dataset, we constructed a corpus from Codeforces blog posts, a popular website for competitive programmers, by extracting its content and user comments. Our results indicate that BTM is powerful enough to do trend analysis in CP. The trend analysis recognized that dynamic programming and complexity analysis have been the most prominent topics for the last ten years. Data structures and string algorithms are runners-up that may have potential trends in the future. This study opens up further research on other methods to perform trend analysis using better topic models and corpora.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 2 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.