I want to conduct a research which is based on time series data ( Suppose, Relation between GDP, Foreign direct investment, Import, Export). Previously I worked with panel dataset and I know the panel dataset's research methodologies properly. Now, I wanted to know about how to do research with time series data.
If someone suggests me Books or articles or any source, where discussed about time series dataset thoroughly, then it will be helpful for me. Thank You.
Md. Maruf Ul Alam @ You could employ the Granger causality test & Johansen's co-integration test to explore the direction of causality & long run relationship between the variables foreign direct investment (FDI) and gross domestic product (GDP). It is also advisable to investigate if bidirectional causality exists between FDI and GDP.
Good luck.
Go to https://z-lib.org/ search for shumway time series analysis with R
And Tsay analysis of financial time series
These are good places to start.
Best, D Booth
Hi, I suggest you the following books :
https://b-ok.asia/book/610800/138576?dsource=recommend
https://b-ok.asia/book/672523/919520
https://b-ok.asia/book/2725596/f6f633?dsource=recommend
https://b-ok.asia/book/610800/138576?dsource=recommend
https://b-ok.asia/book/1183901/9be7ed?dsource=recommend
Thank you Seyed Mahdi Amir Jahanshahi . I will follow your sources.
Hello,
beyond the recommendations by David and Seyed I can recommend the book by Rob Hyndman, inventor of the former forecast package in R
https://otexts.com/fpp2/
(this is an online book)
He and his colleagues/students are heavily involved in the tidyverse approach towards R and his book represents this. The most tremendous difference (in contrast to Shumway and Stoffer and Cowpertwait and Metcliff is the "tsibble" in contrast to the classical ts object. A tsibble (look on youtube for some videos) is a data structure that has an actual time variable (instead of a running index) and several variables (instead of just one as in the ts object). By using time itself, there are functions that detect missing time points and imputes them. And you can apply the rest of useful tidyverse functions (like in dplyr and gglot) to them. And if you need a specific ts object, you can simply create a ts object out of it.
Hope this helps.
--Holger
Hello,
I suggest you to use time series regression if you want to explore the relation between some time dependent variables. Please see following paper:
Article Regression Coefficient and Autoregressive Order Shrinkage an...
- Badges
- Science topic
- Similar topics
- Statistics
- Time Series