Scope of the program

Time series data analysis learning objectives

• Research economic time series for analysis with prediction modeling.
• Analysis of pharmaceuticals producer price index time series (monthly), available from the Federal Reserve Data Base.
• Fundamental theory of time series as discussed in the text “Introduction to Time Series and Forecasting,” by Brokwell and Davis.
• Fitting ARIMA models to the producer price index time series, using differencing to handle non-stationarity.
• Compare ARIMA fits to the time series data analysis over different historical periods.

What challenges did she face?

• During the Mentorship Program, early empirical results suggested a second-order differencing might be necessary to render the time series stationary for analysis. Her efforts determined that the evidence for second-order differencing was due to significant outliers in the data.
• Resolving challenging empirical issues such as this are part of academic research and the student was extremely successful in her resolution.

How did our program help her?

• She would like to work on a challenging project to show on her portfolio for job applications.

Matched mentor in time series data analysis:

The mentor is a professor who teaches financial mathematics and statistics. Through his company, he has been offering consulting services in financial and statistical analytics to a variety of institutions. Citibank, Colonial/Liberty Funds, American Express, and Canon are a few of his former clients. He has been managing investments since 1995, utilizing cutting-edge statistical techniques to oversee a range of investment initiatives.

Recommendation letter

“I am very happy to recommend Roshna to selective universities. In addition to being very bright, Roshna is a delightful person with an optimistic, can-do personality.  I enjoyed every session, from the first when the research topic was developed to the last.“

Lecturer from MIT

Mentor’s evaluation on the student

What was the most difficult aspect of the program the student had to overcome?

Roshna effectively pursued the research agenda throughout all session of the program. Early empirical results suggested a second-order differencing might be necessary to render the time series stationary for analysis. Her efforts determined that the evidence for second-order differencing was due to significant outliers in the data. Resolving challenging empirical issues such as this are part of academic research and Roshna was extremely successful in their resolution.

Do you have any advice for this student as to how the student could further explore this discipline in the future?

The research program referred to chapters of a text on time series data analysis modeling. I encourage her to read further chapters covering advanced topics in this area.

What did you like most about working with this student?

In addition to being very bright, Roshna is a delightful person with an optimistic, can-do personality. I enjoyed every session, from the first when the research topic was developed to the last when detailed statistical estimation results of ARIMA models were presented and compared.

Excerpts from student’s final research paper

Student wrote a final research paper on the topic of ‘Analyzing and Predicting Producer Price Index by Industry: Pharmaceutical Preparation Manufacturing’