Homebrew implementation of IBM Watson DeepQA (NLTK, Semantic Web, AI strategies)
Latest commit 3607d46 Dec 12, 2011 @chriskelvinlee font


A Trivial Pursuit Player A Computer Program Using Natural Language Processing

Michael Chen Christopher Lee Alex McLeese

Computer Science 182 December 12, 2011 Professor Radhika Nagpal

Answering Trivial Pursuit questions

This project provides code that uses natural language processing to answer trivia questions.

Summary of included files

  • determine.py: Given results, determines answer and confidence
  • googleResults.py: Returns the top pages returned by Google, given a query
  • importcache.py: Reads in previously cached results
  • output.py: Caches results
  • questions.py: Encoded triplets of sample questions and answers for testing
  • scoring.py: 4 ways to calculate scores, and a method to use them all
  • test.py: Runs provided questions through given scoring function
  • tp.py: Wrapper to handle import and update
  • trivialpursuitfunctions.py: Given a question and answer options, finds keywords and instances of them
  • weights.py: Functions to determine how heavily each kind of keyword should be weighted