Department of Mathematics & Statistics
Research Abstracts
Research Abstracts
Title: Amazon Electronics Reviews Sentiment Analysis
Authors: Maria Alvarez, Casey Higginson, Cade Lederer, Eleanor Reilly, and Kyle Roberts
Department: Mathematics & Statistics
Abstract: Beginning with a public 2023 dataset containing over 43,000,000 Amazon electronics reviews, this project utilizes the natural language processing technique of sentiment analysis to offer valuable insight into customer attitudes toward the three most commonly reviewed products within a subset of headphones—those from TOZO, Panasonic, and Apple. Using modified versions of the preexisting Bing and AFINN lexicons, as well as a custom stop-word dictionary, we tokenize the reviews and assign sentiments to assess the emotional tone of the words used. We then calculate sentiment scores to classify each review as either positive or negative and isolate reviews that exhibit both a one- or two-star rating and a subzero calculated score. Finally, we investigate these true negative reviews to identify the most pervasive critiques of the earbuds and formulate key takeaways for both producers and consumers.
Title: Breaking Coral: The Sequence
Authors: Elizabeth McCauley, Aiden Rink, and Chris Christensen, Ph.D.
Department: Mathematics & Statistics
Abstract: Coral was a World War II Japanese cipher machine that was in use from 1938 until the end of the war. A machine was never recovered, so what is known about Coral was discovered from codebreaking attacks on the machine. One of the Allied codebreakers' main tasks was to determine the sequence – the letter substitutions – that occurred at the beginning and at the end of the enciphering algorithm. To determine letter substitutions, random samples of enciphered letters were examined based on “roughness,” the Index of Coincidence, and the results of several formulaic approaches. This project analyzes those methods and some modern techniques for determining letter substitutions.
Title: Breaking Jade: The Switches
Authors: Luke Clay and Dr. Chris Christensen
Department: Mathematics & Statistics
Abstract: In 1937, the Japanese designed a trio of machines to encipher messages. One of these machines, called Jade by the Allies, began use in 1941 and was primarily used by naval shore stations. While all three machines used telephone stepping switches, Jade was unique in enciphering katakana and in having five switches. The switches were split between three that stepped and two that were set. The combination of the placement of the switches, the nature of kana frequencies, and the use of only one sequence weakened the overall security of the machine. The attack, carried out by U.S. Navy codebreakers and a group from Bletchley Park in Great Britain, did not last long, and the machine was broken in December 1943. This project describes how the codebreakers recovered the settings of the stepping switches.