Project Folder Structure

• The project has the following folder structure:

  • 00_CODEFORCES_DATA/: Codeforces competitive programming contests collection
  • 01_CODEFORCES_DATASET/: JSON format of the Codeforces competitive programming problem statements-editorials dataset.
  • 02_CODEFORCES_WEBSCRAPPER/: Web scrapper used to fetch all the problem statements and editorials from Codeforces
  • 03_AlgoCorrNet/: Framework used to train/test the proposed system

• Note: To replicate the experiments proposed in the paper 'Matching Problem Statements to Editorials in Competitive Programming' use the framework from 03_AlgoCorrNet

• The datasets referenced in the paper can be found in 03_AlgoCorrNet\Output\Dataset_output

  • CompetitiveProgrammingDataset.csv -> Overall dataset of 1550 competitive programming challanges
  • Preprocessed_CompetitiveProgrammingDataset.csv -> Preprocessed version of the CompetitiveProgrammingDataset
  • Preprocessed_TestingDatasetPath.csv -> Testing Dataset
  • Preprocessed_TrainingDatasetPath.csv -> Training Datset
  • Preprocessed_ValidationDatasetPath.csv -> Validation Dataset
  • CP_MNRLRandomTrainingDataset.csv -> Random partitioned training dataset of the Triple Network Architecture
  • CP_MNRLTrainingDataset.csv -> Unsupervised preprocessed training dataset of the Triple Network Architecture
  • CP_SiameseRandomTrainingDataset.csv -> Random partitioned training dataset of the Siamese Network Architecture
  • CP_SiameseTrainingDataset.csv -> Unsupervised preprocessed training dataset of the Siamese Network Architecture
  • CP_TestingCorrelationDataset.csv -> Pearson / Spearman correlation testing dataset
  • CP_ValidationCorrelationDataset.csv -> Pearson / Spearman correlation validation dataset
  • [email protected] -> Editorial ranking testing dataset
  • [email protected] -> Editorial ranking validation dataset
  • EditorialsIndex.csv -> Collection of 1550 competitive programming editorials