Timeline for How to analyze curvilinear seasonal data
Current License: CC BY-SA 3.0
28 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 26, 2014 at 13:40 | answer | added | goangit | timeline score: 3 | |
| Nov 26, 2014 at 13:06 | comment | added | Aksakal | @rnso this is a panel, except be careful with a term parameter which has a different meaning in statistics | |
| Nov 26, 2014 at 13:00 | comment | added | Nick Cox | I don't follow your distinction at all. Also, please use the word "parameter" in its statistical sense in a statistical forum: there is needless lack of clarity otherwise. You have several subjects followed over time. That's called panel or longitudinal data. Note that @Aksakal raised this point too. There is a lot of reading to do, and this forum contains many pertinent threads. | |
| Nov 26, 2014 at 12:52 | answer | added | IrishStat | timeline score: 1 | |
| Nov 26, 2014 at 11:56 | comment | added | rnso | I appreciate your point. I am not clear about 'panel aspect of data'. From wiki: "Panel data contain observations of multiple phenomena obtained over multiple time periods for the same individuals". However, I am talking about only one parameter of different subjects taken at different times, not repeated testing. | |
| Nov 26, 2014 at 11:51 | comment | added | Nick Cox | There are already many answers here on such methods, so I see no need for another. Here's one: stats.stackexchange.com/questions/60500/… I'll emphasise that from the evidence shown, and what may be guessed, your data may not yield so easily as some physical data to this method. Also, the panel aspect of your data (several people) may need to be addressed. (A major point of the forum is missed if every question is expected to be answered as if it had never been asked before.) | |
| Nov 26, 2014 at 11:44 | comment | added | rnso | @IrishStat : An answer regarding using seasonal ARIMA model here will be very much appreciated. | |
| Nov 26, 2014 at 11:38 | comment | added | Nick Cox | Very easily. Key words range from sinusoids (sines and cosines) through Fourier/periodic/trigonometric regression to TBATS. (Others no doubt.) | |
| Nov 26, 2014 at 10:15 | comment | added | rnso | @NickCox : I am expecting a curvilinear relation. How can use regression for curvilinear data? | |
| Nov 26, 2014 at 9:58 | comment | added | Nick Cox | You don't have to choose either. You can quantify time of year as fraction elapsed since the beginning and with regression methods you don't even need data for every day. Months are just arbitrary units; the body doesn't know it's November. | |
| Nov 26, 2014 at 8:48 | history | edited | rnso | CC BY-SA 3.0 |
edited title
|
| Nov 26, 2014 at 3:17 | history | edited | rnso | CC BY-SA 3.0 |
added 151 characters in body
|
| Nov 26, 2014 at 3:15 | comment | added | rnso | If I do date-wise analysis I will have fewer subjects per day (say 5/day) but with monthly analysis I will have more per month (150/month). Which is better? | |
| Nov 26, 2014 at 3:09 | comment | added | rnso | The question is "Does the variable have a significant seasonal variation?" | |
| Nov 26, 2014 at 3:09 | history | edited | rnso | CC BY-SA 3.0 |
added 78 characters in body
|
| Nov 26, 2014 at 2:58 | comment | added | IrishStat | A seasonal ARIMA model night be appropriate , a deterministic seasonal model with dummies might be appropriate , other models may also be of interest excluding models that use predictors such as time-squared, time-cubed etc. | |
| Nov 26, 2014 at 2:51 | comment | added | IrishStat | For sure you want to understand the method BUT the approach may depend upon the data as @Nick pointed out. The actual data can be used to detail the steps. | |
| Nov 26, 2014 at 2:51 | comment | added | Nick Cox | As @Glen_b hinted, there is no significance assessment without a precise question to be answered. Several quite different models could be fitted here. By the way, a reduction to months could be throwing away useful detail if the data arrive more finely. | |
| Nov 26, 2014 at 2:49 | comment | added | rnso | @IrishStat I want to know the methods which can be used to analyze such data. | |
| Nov 26, 2014 at 2:48 | comment | added | rnso | @NickCox It is a health parameter of many individuals over 3 years. It seems to be seasonality. How can I give a p-value (significance) to it? | |
| Nov 26, 2014 at 2:46 | comment | added | IrishStat | Post the actual numbers not a graph. If you have seasonal data post a number of years . | |
| Nov 26, 2014 at 2:45 | comment | added | Nick Cox | Although a cosmetic detail, try to persuade your software to show integers 1 to 12 as labels, not 2.5, 7.5, 12.5. In the absence of other information, this is a matter of seasonality. But the variability within months should also be unravelled and plotted against date, assuming that you have here the composite of several years' data. The nature of the data may guide appropriate modelling: if it's climatic or climatically driven, then a sinusoid may help; if it's social or economic, then incidence of holidays, vacations, etc. may be important. Half of time series analysis could be relevant.... | |
| Nov 26, 2014 at 2:33 | comment | added | Aksakal | Is it a panel? You can run mixed effects on it if it is. | |
| Nov 26, 2014 at 2:29 | comment | added | rnso | I have added a plot of the data in my question above. | |
| Nov 26, 2014 at 2:28 | history | edited | rnso | CC BY-SA 3.0 |
added 111 characters in body
|
| Nov 26, 2014 at 2:17 | comment | added | Glen_b | "Data" isn't significant or not. What questions do you seek to answer with the data? | |
| Nov 26, 2014 at 2:05 | comment | added | IrishStat | Post your data and perhaps we can help. If your data is .confidential simply scale it. | |
| Nov 26, 2014 at 1:45 | history | asked | rnso | CC BY-SA 3.0 |