patelvap/bashcomplete

Thank you to Saul Greenberg for trace data of Unix users. Citation below.

Greenberg, S. (1988). "Using Unix: collected traces of 168 users". Research Report 88/333/45, includes tar-format cartridge tape. Department of Computer Science, University of Calgary, Alberta.

The graph node structure is contained in graph.py. Parsing functions are defiend in parse.py.

Paper describing methods is located in docs folder.

You may contact [email protected] with questions.

Workflow

The file named predict_n+1_v3 (final).ipynb is the final notebook used for collecting data. You can run this notebook cell by cell to replicate the results. The last run results should be present in the notebook.

The parse.py file will have the relative path of the Unix data as well for parsing and calculating accuracy.

The parser object parses the different user datas with different parsing functions that are named. Terminologies should be defined in the paper.

The cells with print statements are the ones that calculate prediction accuracy and take the most time to run.