PDF Processing and text mining pipeline using Simplicial Complex and Apriori

Authors: Michael Symonds, Samuel Ordonia, Sidharth Mishra, Wei Chung Huang(Wayne Huang)

Advisor: Dr. T.Y Lin

Motive:

• Process and extract PDF(from IEEE Xplore) contents into JSON structures for easy consumption by text-mining applications.

• Analyze the processed PDF documents by passing it through the Simplicial Complex C++ program(old version).

• Write a Simplicial complex version(python wrapper) to work with PDF processor's output(directly); basically construct a data pipline that takes PDF documents, extracts contents out of it into JSON documents, and then consumes the JSON documents to analyze them using our Simplicial complex python wrapper.

• Write the C++ port of Apriori algorithm program included in the Python library "pymining/seqmining" by Barthelemy Dagenais(also called bartdag)

Note: The port was a one-to-one copy of the Python version.

More information about the project layout, contributions and working principles are located in the Team1_CS267_Readme.docx Please refer to the document incase of any concerns.

You are welcome to create issues for this repository to report any issues you run into whilst using this project.

Requirements:

• Python3 (atleast 3.5.2)
• Java 8
• C++ 11 (Visual Studio 2012 or better on Windows)

Note: Python dependecies are included in the requirements.txt. We make use of pdfminer by Yusuke Shinyama for parsing PDFs, after that it is custom logic.