An old project in which I parsed New York Times articles, and created Language Models to recommend the best articles.

Notice: This is meant to provide an example of code. Since I am not allowed to upload source code of my current project, I have chosen this (older) project.

The purpose of this project was to analyze NYT articles (from https://catalog.ldc.upenn.edu/LDC2008T19) and to create language models for recommendation, and then test ways for better, serendipitous recommendations. This project was not finished completely since another approach using topic model yielded better results (publication: A Serendipity Model For News Recommendation, https://hpi.de/naumann/people/maximilian-jenders/year/2015/102852/jenders2015ki.html). Therefore, this project is not complete and not immediately runnable (as you would need the NYT corpus and a database in the back end).

However, this project still shows source code for article stemming and parsing, language model creation, probability calculation, and ranking, in single-threaded and multi-threaded contexts.