BC transfer student becomes Goldwater scholar

Aubrianna Martinez, Senior Digital Editor

Former BC engineering student Annabelle Recinos received the Barry Goldwater Scholarship for her work creating an automated grading assistant. Having transferred from BC to California State University Fullerton in order to complete her bachelor’s degree, Recinos is a civil engineering undergrad and plans on later entering into the Ph.D. program. 

Recinos explained her work on the automated grading assistant that she created with her professor of philosophy Dr. Jonathan Brown and her fellow student Vinh Bui, “through this research project […] I was able to deepen and expand my knowledge of coding. The logic steps that go into coding for a user interface, such as loops and the interface, have helped me greatly in my engineering studies.”

“The model predicts the grade of the new essays based on the input data,” Recinos described via email, thoroughly explaining the process, “Along with the grade, it will also return to the user several of the closest essays to the new essays. The program takes in a large batch of previously graded essays. These essays are then processed by the lemmatizing words, removing stop words are removed and applying the bag of words concept.”

“Each of these three processes helps to put higher weight behind words based on the frequency they appear in the documents and removing common words. The data from the essays is a large array, matrix. To get the core data of the essays the data is passed through a Singular Value Decomposition (SVD) function.”

“From there the data of the new and old essays are compared through the K-Nearest Neighbors method, to return similar essays to the user. Lastly, the software has one last function in which it can return the middle number. It takes into account the frequency of a score, assigning a weight to it based on distance to the new document, and then finds a geometric interpretation of the middle score. This gives a more accurate prediction of scores and is more accurate than just the median.”

Recinos stated that currently only Brown uses the program that they created, and mentioned that he and his team are “still testing out some more novel techniques and will release a final version later for broader distribution.”

Recinos will be using the skills and knowledge she has obtained over the course of her studies to gather research so that she can enhance the transportation industry.