An implementation of frequent pattern mining using the Apriori algorithm to mine meaningful, representative phrases for domains of computer science from titles of academic papers.
The topics are Information Retrieval, Databases, Data Mining, Machine Learning, and Theory. The corpus of paper titles is partitioned using Latent Dirichlet allocation from a natural language processing library.
Simply run 'python mine.py' in the src folder to run the entire program. Check out the results in 'results/pure_patterns' where phrases are ordered from most to least meaningful for each topic!