How to run a nls for multiple dependent variables?

Question

I am unfamiliar with nls however after doing some research on the function it seems to be appropriate for my current analysis.

I am testing population declines over two years of data I have collected from gbif of various species, after taking their total abundance for every month over two years. I wanted to fit a nls following the parameters of this thread here

Here is a subset of my data:

# A tibble: 6 x 5,652
  eventDate   year Charadrius_mongolus Charadrius_alexandrinus Anas_clypeata Anas_penelope
  <date>     <dbl>               <int>                   <int>         <int>         <int>
1 2019-01-01  2019              122244                  133993       2481005       1288829
2 2019-02-01  2019               90248                   45541       1797468        954398
3 2019-03-01  2019               65337                   36938       1874235        740044
4 2019-04-01  2019               40365                   26922       1301156        256811
5 2019-05-01  2019               31051                   18157        289416         53208
6 2019-06-01  2019               11135                   10378         86055         31714

and what I have tried:

nls(Charadrius_mongolus~SSlogis(months, a, b, c), data=mim)

however, I get the following error:

step factor 0.000488281 reduced below 'minFactor' of 0.000976562

I am unsure of what this means and how to approach it.

Furthermore, this ties in with a similar issue I posted in stackoverflow but it may be relevant here also, this is the link

Copy of the dataset:

structure(list(month = 1:30,eventDate = structure(c(17897, 17928, 17956, 17987, 
18017, 18048, 18078, 18109, 18140, 18170, 18201, 18231, 18262, 
18293, 18322, 18353, 18383, 18414, 18444, 18475, 18506, 18536, 
18567, 18597), class = "Date"), year = c(2019, 2019, 2019, 2019, 
2019, 2019, 2019, 2019, 2019, 2019, 2019, 2019, 2020, 2020, 2020, 
2020, 2020, 2020, 2020, 2020, 2020, 2020, 2020, 2020), Charadrius_mongolus = c(122244L, 
90248L, 65337L, 40365L, 31051L, 11135L, 8110L, 32603L, 54299L, 
37389L, 113288L, 82693L, 120867L, 102049L, 54245L, 32205L, 25651L, 
4939L, 9589L, 39085L, 68740L, 65973L, 57434L, 56919L), Charadrius_alexandrinus = c(133993L, 
45541L, 36938L, 26922L, 18157L, 10378L, 17077L, 38381L, 56762L, 
86437L, 62156L, 111735L, 145675L, 45880L, 28012L, 15857L, 11719L, 
9804L, 21872L, 32371L, 100690L, 83973L, 61029L, 75994L), Anas_clypeata = c(2481005L, 
1797468L, 1874235L, 1301156L, 289416L, 86055L, 37646L, 386271L, 
761012L, 1133330L, 1904162L, 1667362L, 2865625L, 2200692L, 2022709L, 
1346010L, 261041L, 70784L, 33624L, 275810L, 1061631L, 1578748L, 
1861672L, 2187776L), Anas_penelope = c(1288829L, 954398L, 740044L, 
256811L, 53208L, 31714L, 14537L, 53650L, 750623L, 1330243L, 904096L, 
1134853L, 1423347L, 953443L, 868277L, 265698L, 65392L, 19920L, 
21885L, 63991L, 879956L, 1585687L, 896546L, 770093L)), row.names = c(NA, 
-24L), class = c("tbl_df", "tbl", "data.frame"))

Can you correct the data frame you have given? It does not contain the variable `month` used in the model! — kjetil b halvorsen, Jul 13 '21 at 18:32
@kjetilbhalvorsen I think that you can reproduce from above the month as I believe its arranged by `eventDate`, so something like: ```mim$month — Stackbeans, Jul 13 '21 at 18:49

score 3 · Answer 1 · answered Jul 14 '21 at 16:22

Good Morning! As I understand you are wanting to fit a logistic curve to your data, however, when plotting your data, I noticed that they are far away from a logistic curve.

par(mfrow = c(1,2))
plot(mim$eventDate, mim$Charadrius_mongolus)
curve(1/(1+exp(-x)), from = -10, to = 10, main = 'Logistic Curve')

As you can see, the curve will not fit your data well. I think another model is better. For example, a polynomial model (I exaggerated a little on the degree of the polynomial, but it is already possible to get an idea of what is happening).

par(mfrow = c(1,1))
plot(1:nrow(mim), mim$Charadrius_mongolus)
model <- lm(Charadrius_mongolus ~ poly(1:nrow(mim), 6), data = mim)
lines(fitted(model), col = 2)

You are right here Jackson! I too had a similar idea after observing these values after a while I may work on using a `GAM` model and fit the data with wiggly splines. Although, were I to use the model you provided - What values could I use from the model to import into a dataframe? — Stackbeans, Jul 14 '21 at 17:17

How to run a nls for multiple dependent variables?

1 Answers1