Timeline for How to estimate uncertainty in Markov chain simulations
Current License: CC BY-SA 4.0
17 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 21, 2020 at 8:15 | vote | accept | MriRo | ||
| Jul 15, 2020 at 14:42 | answer | added | Ryan SY Kwan | timeline score: 2 | |
| Jul 10, 2020 at 12:00 | history | tweeted | twitter.com/StackStats/status/1281558827360563203 | ||
| Jul 10, 2020 at 7:32 | comment | added | MriRo | Thanks! -such a useful platform | |
| Jul 9, 2020 at 20:39 | comment | added | Alexis | Welcome to CV, MriRo! | |
| Jul 9, 2020 at 20:37 | history | edited | Alexis | CC BY-SA 4.0 |
added 5 characters in body; edited title
|
| Jul 9, 2020 at 20:23 | history | edited | kjetil b halvorsen♦ | CC BY-SA 4.0 |
edited tags
|
| Jul 9, 2020 at 14:13 | comment | added | MriRo | I added a paragraph in the question. Please tell me if this made it a little clearer and thanks again for your patience! | |
| Jul 9, 2020 at 14:11 | history | edited | MriRo | CC BY-SA 4.0 |
added 1104 characters in body
|
| Jul 9, 2020 at 13:38 | comment | added | whuber♦ | The source of the uncertainties is clear. What isn't at all clear is what you are doing with them. What exactly are you forecasting and how are you doing it? Could you describe it in English or mathematical notation without requiring readers to understand or run your code? | |
| Jul 9, 2020 at 13:32 | comment | added | MriRo | Sry- There are 2 sources of uncertainty in a random draw of a marcov chain sequence. 1 is based on the probabilities in my transition matrix. These probabilities are estimates for real data so each value has its uncertainty and could be expressed e.g. by a CI. This is the one I asked for in my Q. The other source is the number of random samples I take from my transition matrix to create my sequences and to use for forecasting over time. An infinite number of samples would on average lead to the exact values I expect due to my transition matrix, but a low number (see code) has high uncertainty. | |
| Jul 9, 2020 at 12:53 | comment | added | whuber♦ | You haven't addressed the point: you refer to "simulate the uncertainty due to sample size," but what is the uncertainty of? What is your code intended to estimate? | |
| Jul 9, 2020 at 6:16 | history | edited | MriRo | CC BY-SA 4.0 |
added 21 characters in body
|
| Jul 9, 2020 at 6:13 | comment | added | MriRo | thanks for your help. The quantity is the uncertainty of the estimated values in my transition matrix. This matrix is fitted for data with the markovchainFit() function. But these values are estimates and have uncertainty (see B$$standardError). I want to account for this uncertainty, because this transition matrix is used then to simulate the uncertainty due to sample size in my code. Does this help? e.g. the mean that I plot in the last line would be different when I use slightly different values in my transition matrix B. because the B$estimate is only an estimation with uncertainty. | |
| Jul 8, 2020 at 22:37 | comment | added | whuber♦ | What quantity is it whose uncertainty you wish to estimate? | |
| Jul 8, 2020 at 22:34 | history | migrated | from stackoverflow.com (revisions) | ||
| Jul 8, 2020 at 18:04 | history | asked | MriRo | CC BY-SA 4.0 |