Wolfram Computation Meets Knowledge

Wolfram Summer School

Alumni

Mark Greenberg

Science and Technology

Class of 2019

Bio

Mark Greenberg is a retired public school teacher. Since the 1990s, he has been involved with improving education through the use of computers, particularly video games, for learning. Mark holds a Master’s degree in educational technology and has been using the Wolfram Language for about three years. You can get a glimpse of the work he does at his website, MrGreenberg.com. Mark and his family make their home in sunny Scottsdale, Arizona.

Computational Essay: South Side… Myth?

Project: Poetry Analysis

Goal

Write a program that will take in some poetry (a line, a stanza or a whole poem), analyze it for various features and return a visual representation of its analysis. The first tier will analyze the poem’s meter. I used a neural network and the SequencePredict function to help with this part. The second tier is analysis of rhyme within the poem, a broad definition of rhyme that includes end rhyme, alliteration (front rhyme), consonance and assonance. The purpose of the project is to provide a tool with which students can gain insight into the craft of poetry, something that I believe is sorely needed in schools.

Summary of Results

I found, as I suspected, that machine learning can assist in computer analysis of a natural art form like poetry. Meter and rhyme are physical phenomena that can be measured and described mathematically. Visualizations of various features of poetry are possible and should help students understand what is going on beyond the emotional expression of the human experience we usually present poetry as.

Future Work

Currently, the program makes some errors that I think can be reduced. The neural network that determines words’ phonetic and syllabic representations gives 10% and 4% errors respectively, which might improve with additional training data or parameter adjustment. Also, the scheme for determining the phonetic representation of individual syllables can be improved. After that, I’d like to apply these techniques to analyze the pauses inherent in a written poem and possibly the amount of sentence inversion that a poet uses.