Hi everyone!
I am working on a multi-seasonal time series (hourly) problem. For deseasonalization, I have been using R's mstl package. Initially, I decomposed the time-series into trend, residual, and 2 seasonal components (24 hours and 1 week). I was performing some residual diagnostic to see if there are existing periodicity in the residual by running tests like Ljung-Box test. However, my question is, is it okay to use Ljung-Box test on the residuals delivered by mstl? Because I have always seen these tests being applied on model residuals (not on some decomposed residuals).
Cheers everyone!
I am trying to forecast the Housing Price Index for Toronto and Vancouver for the next 5 years, on Eviews.
My data is monthly and non-stationary, so I ended up taking the second difference for the Toronto data with 24 lags, and the first difference for the Vancouver data, also with 24 lags.
So I applied the Box-Jenkins methodology to estimate models, for Toronto the best fit is ARIMA (1,1,5), and for Vancouver was ARIMA(1,1,1) with an additional MA(3) term.
But while the correlograms of the residuals for both look flat, they don't pass the Ljung-Box test.
Toronto: https://imgur.com/i4ZmGtk Vancouver: https://imgur.com/CfcfOTL
What am I doing wrong? How do I proceed?
I am absolutely not good at this. I'm an A-grade student but this class was just too hard to follow.
Hey,
I've got a stationary time series with around T = 300 observations.I want to test for auto correlation in this series using the Ljung-Box test.
However I'm not sure how to select the number of lags to be included.
I've seen three primary approaches, take min(10, T / 2), ln(T) and take a high value, e.g. 30 and test for all lag number smaller or equal to this value, then take the smallest out of the resulting p-values.
I have two questions:
- Which approach would you take.
- More specifically is the third approach reasonable or will I distort my result by selecting the smallest out of a bunch of theoretically somewhat independent p-values?
I'm happy about every paper I can read / cite on that.
Thanks in advance.
Hi! I'm trying to do a Ljung-Box test to see if the squared residuals meet the GARCH model conditions. I am working with stock returns. I do make my regression:
Datatib$USAVALRet <- log(Datatib$PUSAVAL) - log(lag(Datatib$PUSAVAL))
Then I want to do the LB test:
Box.test(Datatib$USAVALRet, lag = 1, type = c("Ljung-Box"))
But it is not working for some reason... This is the result. It makes no sense considering these are stock prices. Does anybody know why this doesn't work?
X-squared = NaN, df = 1, p-value = NA
I know some family’s have 6-7-8 + people but wouldn’t it be a start at least. No one would be able to hoard necessarily. Then after mailing them all out give them rest away at pharmacies? I mean I get postage would add up but come on we gave away million dollars lottos.
Edit * Temperature control why wouldn’t we send one of those mail pick up tags then to every mailbox so people could use that tag to pick up their box at location x y z. Once every household or dwelling had their box then give the rest away as a free for all.
She waited at U of M walk-in for a little over an hour, was next to be tested and they said she needs to take a rapid test that has to be positive before they will test her. Rest of our family is sick at home and have already been tested and are waiting for results.
As I think about this, it says to me, they are trying to weed out negatives on the PCR to be able to catch up with testing. And honestly handing out the rapid tests will lower the waits.
But if it comes back positive she has to wait in line again for a PCR.
Would be interested to see if they are going walk down the lines at the testing centres handing out tests to the waiting cars
I’ve already seen people trying to bootleg the free ones their kids got from school. My faith in humanity is declining.
The previous thread has kind of washed up. I wanted to start a new one as I have friends who are still looking to get their hands on some at home tests if possible. Please post where you’ve last seen them, thanks!
As the title says, I think they buffet the rates on the next mystery box you open for a limited amount of time. I saw someone who just got a platinum box and I wanted to test it out for myself. I went I opened a box and I got a Gold Box. I opened another one and it was a Silver Box. So I guess that’s why I assume that the buff only works on your next Box that you open. I hope that people can test it out and let me know if that’s actually happening to everyone. I would be excited to see what boxes people get. Also I think is a nice way to say sorry for the inconvenience from the CDC team.
Anyway I am blabbering, fingers crossed that many of you get platinum boxes if that is true. Good luck!
This semester will be hell, I've got 8 people in my family currently positive nearly double that between friends/aquaintances. It's either gonna no differently from last semester or we're gonna be back to zoom university with shitty teachers in no time. I hate to be such an insufferable prick here but this is my last semester, and I hate the idea of being back online and even more at the mercy of teachers who don't care.
Limit 4 per person.
Hey all,
I've seen people on this sub giving stuff away on Tribble (the test server, not the actual live/holodeck server), and since one big point of contention for some players is not being able to test lobi or lockbox stuff if they don't have master keys on live, I thought I'd help out in a small way with some offers.
You can only pick one of the offers below, once per account. This restriction is put in place to keep the workload fair for everyone, as well as giving everyone who participates a higher chance of getting what they ask for.
Master key offer:
If you can't copy master keys over to Tribble from the live server, but you want to test lock box stuff on TRIBBLE, put in a request as per below and I'll try and get you some master keys on Tribble to play with there. While the exact number of master keys given may be different, it will be reasonable enough to get a good go at lock boxes (at least 10, maybe more if its reasonable for me to do so).
Item offer:
If you want a specific lock box (Infinity or Discovery only) or lobi ship item pack on TRIBBLE (I can only do kits, traits, weapons packs and a limited [a few of the first requests at my discretion] amount of Lobi or Infinity pack T5 ships) and I have it already or can get it without too much fuss, I will try and send the item to you - if you want this option, put in a request as per below and I will mail you the item if this is the case.
How to request the offers:
Please use the below post format please (use bracketed guidance, and delete the bracketed stuff in your actual post). Posts not in this format may not be able to be processed.
| Handle: | (your STO handle) | |
|---|---|---|
| Offer(s) requested: | (write 'master key' or 'item') | |
| (only if an item is requested) Item: | (in choice FROM pack name order - e.g. Need-to-Know Basis FROM Infinity Prize Pack - Starship Trait) |
| (write "Discovery LB item" if you are requesting a Discovery Lockbox item, otherwise write nothing) |
|---|
| (write "Lobi item" if you are requesting a Lobi Store ship, otherwise write nothing) |
Happy testing!
Wondering if the Gom Jabbar Box would only be effective the first time on the user?
If it was administered a second time, the user would know that the pain is only “imagined” and temporary, and it might not have the same weight.
Any knowledge or lore on this?
Multikonstnären Lasse Åberg (tigern) tillsammans med gitarrlegenden Janne Schaffer (zebran) och pianisten Peter Ljung (pantern) delar med sig av sina helt makalösa karriärer som sträcker sig nära 60 år i konstens och musikens tecken.
#artyria #janneschaffer #lasseåberg #peterljung
https://bananklubben.se/aberg-schaffer-och-ljung-inviger-artyria/
I shipped mine on the 11th and have yet to receive any result. The wording on what was allowed on in the drop boxes was a little confusing and now I’m not sure whether they actually processed it or not. Did anyone else use a usps drop box and did you get a result back?
Correct typo - disconnected* /discounted
