About

I grew up in a small town in the Shandong province, which is located in central eastern China. When I was thirteen, My father and I moved to Athens, GA, following my mother who was pursuing her Ph.D. degree at the University of Georgia. I have since then lived in Athens, completing middle school, high school, college, and Ph.D. I also have a little brother who is fourteen years younger than me. He currently lives with our parents in Virginia.

Research

My general research interest is in Data Science, and in particular, Forecasting using both Statistical Models and Machine Learning Models, developing theory-guided data science models for forecasting, with applications in Traffic, Economics and Finance.

Forecasting

Forecasting is an area of interest in many people's lives because the ability to anticipate the future can be very beneficial in terms of making investment decisions, travel plans, etc. The presence of Big Data has also enabled data scientists to test, validate and evaluate forecasting models on increasing amounts of data, which may lead to greater forecasting accuracies. Various forecasting models have been developed, and may be generally divided into Statistical Models and Machine Learning Models.

  • Statistical Models

    Classical forecasting models include ARIMA and Exponential Smoothing, for which I have implemented in the ScalaTion Project and evaluated for traffic flow forecasting. Other statistical forecasting models include GARCH, commonly used to model financial volatility; the VARMA family of models, which are multivariate generalizations of univariate ARIMA family; Functional AutoRegressive (FAR) model, which is used to model functional time series; Hidden Markov Models and State Space Models, which have been used for a great number of applications including econometrics, temporal pattern recognitions such as speech and handwriting, among others. The Statistical Models in general are based on well established theories, assumptions and equations; in contrast to the recent trends of Machine Learning Models, which are much more data-driven and much less emphasis are placed on the interpretability of the learned parameters. I am interested in learning and evaluating both Statistical Forecasting Models and Machine Learning Forecasting Models for various types of time series from the fields of Traffic, Economics and Finance.

  • Machine Learning Models

    On the Machine Learning's side, techniques such as Support Vector Regression and Neural Networks are commonly used for forecasting. Neural Networks in particular have garnered much attention due to recent progress made in deep learning for image classification problems. Various types of Neural Networks exist, including the standard fully-connected multi-layered feedforward Neural Networks, for which I have provided an implementation in ScalaTion; Convolutional Neural Networks, which are able to take advantage of spatial dependencies in the data; Recurrent, and in particular, Long Short-Term Memory Neural Networks, which are able to take advantage of the temporal dependencies in the data, a much needed characteristic for analyzing time series.

Traffic

In recent years, large numbers of traffic sensors have been deployed throughout major urban areas and freeways in the United States. These sensors collect great volumes of data such as traffic flow, speed, lane occupancy, etc. The ability to produce accurate traffic forecasts can lead to smarter traffic management and control systems (i.e., smart traffic lights) as well as the development/improvements of map applications with better route selections for advanced trip planning. There have also been well established, theory-based, traffic simulation models such as microscopic and macroscopic traffic flow models. The recent trends in Data Science research place great values on the usage of data to learn the parameters of a model, but the process is often like a black box. Well established theories should be used to help guide and constrain the learning process, yet there has not been enough attention placed on this process of creating theory-guided data science models, for which I am interested in devoting further research in.

Economics and Finance

Forecasting has been the central theme of research in the fields of Economics and Finance. The benefits of producing accurate forecasts for financial time series are obvious in terms of investments. The ability to anticipate the state of the economy may help administrations to respond with appropriate policies. Much data are freely available in those fields, from the FRED database that contains more than five hundred thousand time series to the historical and current prices of stocks, cryptocurrency, real estate values, among many others. A plethora of economic and financial models have also been developed, from simple and intuitive mathematical equations and graphs to more complex ones that may required advanced numerical methods such as differential equations. Researchers should be able to take advantage of these theories and construct models that are both coherent with the theories and optimized using large amounts of available data in order to produce better forecasts.

CV

A pdf version of my CV can be download here.

Contact

  • Email: penghga@uga.edu
  • Office Location: Boyd 614

Elements

Text

This is bold and this is strong. This is italic and this is emphasized. This is superscript text and this is subscript text. This is underlined and this is code: for (;;) { ... }. Finally, this is a link.


Heading Level 2

Heading Level 3

Heading Level 4

Heading Level 5
Heading Level 6

Blockquote

Fringilla nisl. Donec accumsan interdum nisi, quis tincidunt felis sagittis eget tempus euismod. Vestibulum ante ipsum primis in faucibus vestibulum. Blandit adipiscing eu felis iaculis volutpat ac adipiscing accumsan faucibus. Vestibulum ante ipsum primis in faucibus lorem ipsum dolor sit amet nullam adipiscing eu felis.

Preformatted

i = 0;

while (!deck.isInOrder()) {
    print 'Iteration ' + i;
    deck.shuffle();
    i++;
}

print 'It took ' + i + ' iterations to sort the deck.';

Lists

Unordered

  • Dolor pulvinar etiam.
  • Sagittis adipiscing.
  • Felis enim feugiat.

Alternate

  • Dolor pulvinar etiam.
  • Sagittis adipiscing.
  • Felis enim feugiat.

Ordered

  1. Dolor pulvinar etiam.
  2. Etiam vel felis viverra.
  3. Felis enim feugiat.
  4. Dolor pulvinar etiam.
  5. Etiam vel felis lorem.
  6. Felis enim et feugiat.

Icons

Actions

Table

Default

Name Description Price
Item One Ante turpis integer aliquet porttitor. 29.99
Item Two Vis ac commodo adipiscing arcu aliquet. 19.99
Item Three Morbi faucibus arcu accumsan lorem. 29.99
Item Four Vitae integer tempus condimentum. 19.99
Item Five Ante turpis integer aliquet porttitor. 29.99
100.00

Alternate

Name Description Price
Item One Ante turpis integer aliquet porttitor. 29.99
Item Two Vis ac commodo adipiscing arcu aliquet. 19.99
Item Three Morbi faucibus arcu accumsan lorem. 29.99
Item Four Vitae integer tempus condimentum. 19.99
Item Five Ante turpis integer aliquet porttitor. 29.99
100.00

Buttons

  • Disabled
  • Disabled

Form