Week 12 – A Successful End

May 05, 2019

Success! After presenting my first presentation to my advisor and a class, I got some good feedback to improve my slides–I feel prepared for the upcoming Senior Project Committee presentations, between the new advice and additional practice.

Beyond my presentation, however, I was actually able to bring my method’s accuracy up from a little over 50% to 74.34%, a significant improvement to my project. My method, which takes an AN phrase and compares the concreteness ratings of the adjective and the noun, determined whether or not a phrase was metaphorical by seeing if the concreteness ratings were within 1.2 of one another. I’ve changed it, however, to be a bit more intuitive–it now follows a set of “rules,” listed below:

Abstract adjectives with concrete nouns tend to be metaphorical

Abstract adjectives and abstract nouns tend to be metaphorical

Concrete adjectives with concrete nouns tend to be concrete

Concrete adjectives and abstract nouns tend to be metaphorical

The code itself is more complicated, but it acts along these lines. After testing, here’s a screenshot of a test run of the new method:

To give a frame of reference, Peter D. Turney, who I based my work off of, achieved an accuracy of 79%. I believe that this difference in accuracy can be attributed to the datasets–Turney used machine learning to generate his own, custom dataset. I’m very proud of the result of my project!

Leave a Reply

Your email address will not be published.