William W.S. Wei, PhD, is a Professor of Statistics at Temple University in Philadelphia, Pennsylvania, USA. He has been a Visiting Professor at many universities including Nankai University in China, National University of Colombia in Colombia, Korea University in Korea, National Chiao Tung University, National Sun Yat-Sen University, and National Taiwan University in Taiwan, and Middle East Technical University in Turkey. His research interests include time series analysis, forecasting methods, high dimensional problems, statistical modeling, and their applications.
Wei’s latest book, Multivariate Time Series Analysis and Applications, has recently been published by John Wiley and Sons Ltd and is a part of the Wiley Series in Probability and Statistics. After the success of Time Series Analysis—Univariate and Multivariate Methods, Wei’s latest text focuses on high dimensional multivariate time series, and is illustrated with numerous high dimensional empirical time series. Due to the development of computers and the internet, we have increased potential for data exploration. In the next few years, dimension will become a more serious problem. The author provides some initial solutions, which may encourage the development of related software needed for the high dimensional multivariate time series analysis.
To learn more about this recently publication, Fran McMahon spoke with William W. S. Wei about the book and the work behind it. Read on to learn more.
1. Congratulations upon the publication of Multivariate Time Series Analysis and Applications, an essential guide on high dimensional multivariate time series. What led to the idea to writing this book?
My first time series analysis book, Time Series Analysis – Univariate and Multivariate Methods, has been used by many researchers and universities worldwide. Because of the development of high speed devices and technology, we have had a data explosion. New theories and methods have been developed for high dimensional time series analysis. Many publishers have contacted me for a new edition. Because of the development of so much new material, it would be impractical to include them all in a new edition of the book, so I decided to write a new book.
2. As it is noted, this book is not an introduction to this field of study, but an advanced analysis. However, for someone who may not be well versed in this subject area, how would you explain the topic and study to them?
In our daily life, we often encounter time series data, which is simply a data set related to time. For example, daily stock prices of the Dow Jones Industrial Average and S&P 500. Statistical theory and methods to analyse these time series is called time series analysis. When we concentrate on the analysis of a single time series, we use univariate time series analysis. When we extend our analysis to several related time series, we use multivariate time series analysis, which often involves high dimensional issues in the analysis.
3. What were your main objectives during the writing process? And which topic, that you discuss in your book, would you say was the most interesting studying?
My main objectives were to have a good introduction to multivariate time series, a comprehensive collection of theory and methods for analysing these series, and new development of high dimensional multivariate time series analysis. The last part is clearly the most interesting to study, where I introduce new multivariate spectral analysis methods, including the analysis of nonstational multivariate time series and various dimension reduction methods for high dimensional multivariate time series.
4. If there is one piece of information or advice that you would want your audience to take away from your book, what would that be?
Enjoy the read and try some methods on your multivariate time series.
5. Who should read the book and why?
I strongly encourage readers and researchers who are using multivariate time series in their study to read this book because it includes many topics that have not been introduced in other time series analysis books.
6. Why, do you think, this area of study may be of interest now?
As I have indicated earlier, because of the development of rapid growing technology, we have had a big data explosion. Researchers are interested in and working on new theories and methods for high dimensional data sets daily.
7. Alongside your own text, what other books would you recommend to students looking to learn more about high dimensional multivariate time series analysis?
To students who are looking to learn more about high dimensional multivariate time series analysis, I strongly encourage them to search for new papers on related topics via the internet. There are several books on multivariate time series which readers can easily find through websites. However, at this point, in terms of a high dimensional multivariate time series analysis book, I do not have any recommendation.
8. What is it about the area of high dimensional multivariate time series analysis that fascinates you?
With advancements in technology, we have large amounts of data in numerous fields such as business trading, genomics, and images. Developing new methodology to analyze this data is fascinating me. As a statistician, I want to know the best inference method for the high dimensional big data.
9. What other work are you currently working on or has recently been published?
There are many methods available to analyse multivariate time series. I am currently working on comparisons of various methods in terms of their estimations and forecasting.