Possible bug with subfits when mppm formula involves splines

[agila5]

Dear spatstat developers, I created this issue to report a problematic interaction between spatstat.model and splines when the underlying mppm object was defined using a natural cubic spline. For example:

library(spatstat)
library(splines)
H <- hyperframe(Y = waterstriders)
obj <- mppm(Y ~ ns(x, df = 3),  data=H)
subfits(obj)
#> Error in `ns()`: ! all interior knots match left boundary knot

^{Created on 2026-02-05 with reprex v2.1.1.9000}

After some simple debugging, I noticed that the error occurs in spatstat.model::impliedpresence() when running model.frame(). Based on my understanding, it seems that impliedpresence() tries to build a "fake" model matrix row-by-row starting from a matrix of all zeros. Moreover, I run a few more tests and noticed that exactly the same error occurs when running analogous code outside of spatstat, for example:

library(splines)
toy <- data.frame(x = 0)
model.frame(~ ns(x, df = 3), data = toy)
#> Error in `ns()`: ! all interior knots match left boundary knot

^{Created on 2026-02-05 with reprex v2.1.1.9000}

Do you have any suggestion on how to proceed? Thanks!

Session info

sessioninfo::session_info()
#> Warning in system2("quarto", "-V", stdout = TRUE, env = paste0("TMPDIR=", :
#> running command '"quarto"
#> TMPDIR=C:/Users/andre/AppData/Local/Temp/Rtmp6xPOc8/file55d0bb173aa -V' had
#> status 1
#> ─ Session info ───────────────────────────────────────────────────────────────
#>  setting  value
#>  version  R version 4.4.2 (2024-10-31 ucrt)
#>  os       Windows 11 x64 (build 26200)
#>  system   x86_64, mingw32
#>  ui       RTerm
#>  language (EN)
#>  collate  Italian_Italy.utf8
#>  ctype    Italian_Italy.utf8
#>  tz       Europe/Rome
#>  date     2026-02-05
#>  pandoc   3.2 @ C:/Program Files/RStudio/resources/app/bin/quarto/bin/tools/ (via rmarkdown)
#>  quarto   NA @ C:\\Users\\andre\\AppData\\Local\\Programs\\Quarto\\bin\\quarto.exe
#> 
#> ─ Packages ───────────────────────────────────────────────────────────────────
#>  package          * version    date (UTC) lib source
#>  abind              1.4-8      2024-09-12 [1] CRAN (R 4.4.2)
#>  cli                3.6.5      2025-04-23 [1] CRAN (R 4.4.2)
#>  deldir             2.0-4      2024-02-28 [1] CRAN (R 4.4.2)
#>  digest             0.6.37     2024-08-19 [1] CRAN (R 4.4.2)
#>  evaluate           1.0.3      2025-01-10 [1] CRAN (R 4.4.2)
#>  fastmap            1.2.0      2024-05-15 [1] CRAN (R 4.4.2)
#>  fs                 1.6.5      2024-10-30 [1] CRAN (R 4.4.2)
#>  glue               1.8.0      2024-09-30 [1] CRAN (R 4.4.2)
#>  goftest            1.2-3      2021-10-07 [1] CRAN (R 4.4.2)
#>  htmltools          0.5.8.1    2024-04-04 [1] CRAN (R 4.4.2)
#>  knitr              1.50.4     2025-07-11 [1] https://yihui.r-universe.dev (R 4.4.2)
#>  lattice            0.22-6     2024-03-20 [2] CRAN (R 4.4.2)
#>  lifecycle          1.0.4      2023-11-07 [1] CRAN (R 4.4.2)
#>  Matrix             1.7-1      2024-10-18 [2] CRAN (R 4.4.2)
#>  mgcv               1.9-3      2025-04-04 [1] CRAN (R 4.4.2)
#>  nlme             * 3.1-166    2024-08-14 [2] CRAN (R 4.4.2)
#>  pillar             1.10.2     2025-04-05 [1] CRAN (R 4.4.3)
#>  polyclip           1.10-7     2024-07-23 [1] CRAN (R 4.4.2)
#>  reprex             2.1.1.9000 2025-02-25 [1] Github (tidyverse/reprex@07cd5d7)
#>  rlang              1.1.6      2025-04-11 [1] CRAN (R 4.4.3)
#>  rmarkdown          2.29       2024-11-04 [1] CRAN (R 4.4.2)
#>  rpart            * 4.1.23     2023-12-05 [2] CRAN (R 4.4.2)
#>  rstudioapi         0.17.1     2024-10-22 [1] CRAN (R 4.4.2)
#>  sessioninfo        1.2.3      2025-02-05 [1] CRAN (R 4.4.2)
#>  spatstat         * 3.5-1      2026-02-01 [1] https://spatstat.r-universe.dev (R 4.4.2)
#>  spatstat.data    * 3.1-9      2025-10-18 [1] https://spatstat.r-universe.dev (R 4.4.2)
#>  spatstat.explore * 3.7-0.002  2026-02-03 [1] https://spatstat.r-universe.dev (R 4.4.2)
#>  spatstat.geom    * 3.7-0      2026-01-20 [1] https://spatstat.r-universe.dev (R 4.4.2)
#>  spatstat.linnet  * 3.4-1      2026-01-31 [1] https://spatstat.r-universe.dev (R 4.4.2)
#>  spatstat.model   * 3.6-1      2026-01-29 [1] https://spatstat.r-universe.dev (R 4.4.2)
#>  spatstat.random  * 3.4-4      2026-01-21 [1] https://spatstat.r-universe.dev (R 4.4.2)
#>  spatstat.sparse    3.1-0      2024-06-21 [1] CRAN (R 4.4.2)
#>  spatstat.univar  * 3.1-6.002  2026-02-03 [1] https://spatstat.r-universe.dev (R 4.4.2)
#>  spatstat.utils     3.2-1      2026-01-08 [1] https://spatstat.r-universe.dev (R 4.4.2)
#>  tensor             1.5        2012-05-05 [1] CRAN (R 4.4.2)
#>  vctrs              0.6.5      2023-12-01 [1] CRAN (R 4.4.2)
#>  withr              3.0.2      2024-10-28 [1] CRAN (R 4.4.2)
#>  xfun               0.52       2025-04-02 [1] CRAN (R 4.4.2)
#>  yaml               2.3.10     2024-07-26 [1] CRAN (R 4.4.2)
#> 
#>  [1] C:/Users/andre/AppData/Local/R/win-library/4.4
#>  [2] C:/Program Files/R/R-4.4.2/library
#>  * ── Packages attached to the search path.
#> 
#> ──────────────────────────────────────────────────────────────────────────────

[baddstats]

Thank you for this information. (In the example given, the error arises in impliedcoefficients rather than impliedpresence, but both these functions cause similar errors.)

This is a weakness of ns and bs (related to the problem of unsafe prediction). These functions create a basis that depends on the data, so prediction is not straightforward.

In the short term, it is unlikely that we will be able to find a workaround.

You could try adding the argument Boundary.knots to the call to ns which would at least remove the error.
Example

H <- hyperframe(Y = waterstriders)
xr <- c(-1, 50)
obj <- mppm(Y ~ ns(x, df = 3, Boundary.knots=xr),  data=H)
subfits(obj)

However, prediction from the subfits may still be unreliable.