report sparsedist

[vbartolino]

I do not entirely understand the output from fit$sparsedist
The scale in model_n is different from what I would expect based on obs_mean
Am I thinking wrong?

> fit$sparsedist[1:5,] %>% data.frame()
  year step area data_type function_f component age length obs_mean obs_stddev
1 2008    2 <NA>   asparse     linreg   bt2.rin   4  117.0      1.5         NA
2 2008    2 <NA>   asparse     linreg   bt2.rin  12  121.5      1.5         NA
3 2008    2 <NA>   asparse     linreg   bt2.rin  13  126.5      2.3         NA
4 2008    2 <NA>   asparse     linreg   bt2.rin   5  131.0      1.8         NA
5 2008    2 <NA>   asparse     linreg   bt2.rin   7  131.0      2.0         NA
  obs_n model_sum model_sqsum  model_n
1    NA 177.98198    411.0834 77.05877
2    NA  68.27062    168.8585 27.60227
3    NA  60.21206    148.9279 24.34395
4    NA 104.69718    192.3725 56.98058
5    NA 119.86707    268.4067 53.53113

[lentinj]

If a fleet/predator isn't provided (which I don't think you do), then model_n should be the sum of abundance for that year/step/age/length. If fleet/predators are provided, then its the indivduals caught for that year/step/age/length.

There's a missing column, model_mean = model_sum / model_n, that should be comparable to obs_mean. Generating this in g3_fit seems a fairly simple addition.

[willbutler42]

Hi, fyi, I have added the model_mean column to the output: 59fd1cb.

[lentinj]

@willbutler42 thanks!

[vbartolino]

I noticed a minor issues with this report when I include two distinct sparsedist likelihood components, in my case
(1) individual BT 72 observations (bt1.rin to capture seasonality)
(2) a time series of 24 annual BT (bt2.rin to capture long term trends)

The problem appears when (1) has a name which makes it listed after (2) in fit$sparsedist.

Here the result

individual data bt1.rin and time series bt2.rin OK

> fit$sparsedist %>% filter(component == "bt1.rin") %>% select(year,step,age,obs_mean,model_mean) %>% head(25) %>% data.frame()
   year step age obs_mean model_mean
1  2008    2   4      1.5   1.868525
2  2008    2  12      1.5   2.015113
3  2008    2  13      2.3   2.015124
4  2008    2   5      1.8   1.583205
5  2008    2   7      2.0   1.870232
6  2008    2   6      1.4   1.778922
7  2008    4   8      5.3   4.595219
8  2008    4  17      5.0   4.523513
9  2008    4   7      6.5   4.456400
10 2008    4  11      5.4   4.431778
11 2009    4  12      6.5   4.619059
12 2011    4  23      6.7   4.777515
13 2011    2   3      6.0   2.710815
14 2012    3   6      4.0   3.256975
15 2012    2   8      1.8   2.146811
16 2013    3   3      4.0   3.840378
17 2013    2  20      1.3   2.126861
18 2013    3  11      2.6   3.527726
19 2014    2  11      2.5   2.237870
20 2015    2  13      2.5   2.257318
21 2015    3   6      4.9   3.428736
22 2015    4   3      4.8   4.736945
23 2015    4   3      5.4   4.792316
24 2015    3   9      5.0   3.758631
25 2015    4  11      5.4   5.194392

individual data bt2.rin and time series bt1.rin (the output gets odd)

> fit3$sparsedist %>% filter(component == "bt2.rin") %>% select(year,step,age,obs_mean,model_mean) %>% head(25) %>% data.frame()
   year step age obs_mean model_mean
1    NA    2   4       NA         NA
2    NA    2  12       NA         NA
3    NA    2  13       NA         NA
4    NA    2   5       NA         NA
5    NA    2   7       NA         NA
6    NA    2   6       NA         NA
7    NA    4   8       NA         NA
8    NA    4  17       NA         NA
9    NA    4   7       NA         NA
10   NA    4  11       NA         NA
11   NA    4  12       NA         NA
12   NA    4  23       NA         NA
13   NA    2   3       NA         NA
14   NA    3   6       NA         NA
15   NA    2   8       NA         NA
16   NA    3   3       NA         NA
17   NA    2  20       NA         NA
18   NA    3  11       NA         NA
19 2008    2  11      1.5   2.276591
20 2008    2  13      1.5   2.213743
21 2008    3   6      2.3   2.213783
22 2008    4   3      1.8   1.478098
23 2008    4   3      2.0   1.884802
24 2008    3   9      1.4   1.843897
25 2008    4  11      5.3   4.535867