Timeline for Understand order of time series
Current License: CC BY-SA 3.0
13 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 23, 2015 at 18:16 | comment | added | Learner | Thanks a lot Dave for the clarification. Your inputs have been very helpful in my understanding of this data set and time series analysis in general. I will get in touch offline in case I have further queries | |
| Jul 22, 2015 at 22:12 | history | edited | IrishStat | CC BY-SA 3.0 |
added 1527 characters in body
|
| Jul 22, 2015 at 21:10 | comment | added | Learner | Also I am facing some difficulties in understanding the model equation from autobox 1) Is level shift a dummy variable? For example is the score of level shift 3 (1792) is 0 for data points before 1792? 2) How does historical pulse affect current observation? Is it through impulse-response? 3) How to read the order of pulses. X5,X6,X7 refer to pulses 515,702,514 why is the order not sequential? 4) How to read [(1-0.348B1)]-1 [A(T)] | |
| Jul 22, 2015 at 21:10 | comment | added | Learner | However for the GLM model I need to estimate the covariance matrix and I am not sure how can I estimate the matrix. For example a model with lag 1 will require a 2*2 matrix and If I use the errors for covariance estimation, the diagonal elements will be more or less same and the off diagonal elements will depend only on the auto correlation. Hence this will only address auto correlation and not heteroscedasticity | |
| Jul 22, 2015 at 21:09 | comment | added | Learner | I have come to the understanding that I will have to use an iterative method for parameter estimation. As of now I am considering the following approach: 1) Find out the level shifts using dummy variables 2) Fit a model using level dummy and AR(1) 3) Using the residuals of the above models identify pulses (I believe they are not seasonal) 4) Build model using both level and pulses and apply GLM to the residuals to find AR estimates However for the GLM model I need to estimate the covariance matrix and I am not sure how can I estimate the matrix. | |
| Jul 22, 2015 at 21:08 | comment | added | Learner | Hi Dave, actually I have few queries and since your answers I had been doing research on how best I can implement your inputs using the available statistical software . It will be of great help if you could guide me in the following points. I have divided my comment into multiple comments as it's too long for the comment textbox | |
| Jul 22, 2015 at 15:53 | comment | added | IrishStat | @Learner Is my answer clear. If not please feel free to contact me and I will try and verbally communicate with you (dialogue) | |
| Jul 21, 2015 at 11:45 | history | edited | IrishStat | CC BY-SA 3.0 |
added 262 characters in body
|
| Jul 21, 2015 at 10:31 | comment | added | IrishStat | See my EDIT which responds to your comment . | |
| Jul 21, 2015 at 10:30 | history | edited | IrishStat | CC BY-SA 3.0 |
added 714 characters in body
|
| Jul 21, 2015 at 8:50 | vote | accept | Learner | ||
| Jul 21, 2015 at 8:50 | comment | added | Learner | Hi Dave, thanks a lot for the detailed explanation. I am currently trying to understand in more details about the points you mentioned in your comment. Could you please provide some details about how to does the x axis representation of the above plot relate to seasonality of 7? And which transformation is applied to the y axis? | |
| Jul 19, 2015 at 19:03 | history | answered | IrishStat | CC BY-SA 3.0 |