README

This is a simple spam email classifier, which I use to showcase a data analysis pipeline.

How to reproduce this analysis

• Clone the repository!
• Place the test emails in data/ and follow the naming and format specification below.
• Run make clean to cleanup workspace from previous analysis.
• Run make INPATH=data/testemails.SAMPLE_ID.txt, replacing SAMPLE_ID to match with the path to file containing emails for testing.
• Classification results will be found in results/ for files with name prediction.CLS_ID.SAMPLE_ID.csv (where CLS_ID is the ID of the classifier used).

test email format specification

• It should follow the same format as [the sample](
• the nmake INPATH=data/testemails.fake.txt

Autotest

It is adviced to leave the auto-test running while making changes. At project root, run: Rscript tests/runtests.R which runs all tests in the tests/ directory first, and rerun a test if a change in either the test or the code in spClassR/ is detected.

Dependencies

• GNU make tool
• Python
• R
• R packages
  • tm
  • plyr
  • kernlab
  • randomForest
  • e1071
  • testthat