1

I have two sets of time series data - roleTrajectories & normalizedDegree. The former data set contains ordinal rankings of subjects' positions within a network at 13 time periods. The latter data set contains subjects' normalized degree centrality at the same 13 time periods. I am trying to determine whether or not subjects' changes in ordinal rank is correlated with changes in their normalized degree centrality.

What is most appropriate approach to analyzing the relationship between the two data sets?

I plan on using R for the analysis; if possible I would be grateful if someone could point me to the right package to use.

Mark C
  • 85
  • 8
  • "How do I analyze a relationship" is not usually a question well answered identifying "which test". That only answers a very specific kind of question (and often not a very revealing one). Considering what your questions of interest are, and what a suitable model for the data could be might be a good place to start. Then - if necessary - what tests will answer questions of interest might become clearer. – Glen_b Jan 04 '15 at 01:24

1 Answers1

1

Here are some introductory guides on performing time series analysis in R:

If you're interested in more detailed exploration of the rich variety of R packages, focused on time series analysis, and their functionality, the corresponding CRAN Task View: Time Series Analysis is IMHO the best place to start. Correspondingly, an excellent and one of the most comprehensive R-focused guides on the topic is the "Time Series Analysis with R" by MacLeod, Yu and Mahdi.

If you're interested in forecasting in regard to your time series data, the following free and also R-based online textbook "Forecasting: Principles and practice" by Hyndman and Athanasopoulos could be very helpful: https://www.otexts.org/fpp.

Bonus / Fun fact: The 2003 Nobel Prize in Economics has been awarded to professors Robert Engle and Clive Granger for their work on economic time series analysis (volatility and trends, correspondingly). Research paper, summarizing their work and implications can be found here.

Aleksandr Blekh
  • 7,867
  • 2
  • 27
  • 93