BYU

Abstract by Xin Zhao

Personal Infomation


Presenter's Name

Xin Zhao

Degree Level

Undergraduate

Abstract Infomation


Department

Physics and Astronomy

Faculty Advisor

Mark Transtrum

Title

Crowd Noise Data-piping System

Abstract

The goal of our research is to obtain an accurate machine learning model to identify or predict crowd nosie state in a sport event . This is a challenging process due to the great variety of data and the flexibility of machine learning training methods that are involved in it. Thus, we are developing a data piping system that can used to manage the training data and execute certain training methods to help simplify and unify this process within the research group from past year. The specific functions of the data-piping system are divided into raw acoustic data conversion, feature data classification and storage, and efficient data acquisition. By doing so, all the research related data will be available online with proper authentication from BYU network. And we can save plenty of time in the process of changing raw acoustic data to functional trained machine learning models trough a simple client script.