A Mathematical Study of Bourbon and Other Alcohols

Brianna Adler Thomas More College, Isabella Miller Thomas More College
Faculty Sponsor(s): Jyoti Saraswat Thomas More College, William Wetzel Thomas More College
In this research study, we look at the mathematical analysis of the chemical fingerprints of different alcohols including bourbon, whiskey, tequila, brandy, rum, gin and liqueur. The data collected over the span of couple of years involves gas chromatography-mass spectrometry (GC-MS) to separate and identify the chemical compounds at the molecular level. This dataset, which primarily consists of chemical compounds, is analyzed mathematically for patterns. Building upon previous work, we proceed with the task of pattern recognition using different unsupervised techniques. This study focuses on cluster analysis, which is to group our data into subsets called clusters. Our go to approach for this study is the K-means algorithm. Unlike other cluster approaches, K-means is a top-down procedure. The problem of how to choose the number of clusters remains a big challenge. The numerical results of K-means obtained with Euclidean distance indicate the clusters are not well separated, but in fact segmented.
Mathematics & Computer Science
Oral Presentation

When & Where

01:30 PM
Gallahue Hall 105