Feature/issue-3107 add beta negative binomial lpmf

[lingium]

Summary

With this PR the lpmf of beta negative binomial distribution are added.
See issue #3107

Expressions involved:

PMF

$$f(n|r, \alpha ,\beta)={\frac {\Gamma (r+n)}{n!\Gamma (r)}}{\frac {\Gamma (\alpha +r)\Gamma (\beta +n)\Gamma (\alpha +\beta )}{\Gamma (\alpha +r+\beta +n)\Gamma (\alpha )\Gamma (\beta )}}$$

Partial derivatives w.r.t the three parameters

$$ \begin{aligned} \frac{\partial \log f}{\partial r} &= \psi(n+r) - \psi(n+r+\alpha+\beta) - \psi(r) + \psi(r+\alpha) \\ \frac{\partial \log f}{\partial \alpha} &= \psi(\alpha+\beta) - \psi(n+r+\alpha+\beta) - \psi(\alpha) + \psi(r+\alpha) \\ \frac{\partial \log f}{\partial \beta} &= \psi(\alpha+\beta) - \psi(n+r+\alpha+\beta) + \psi(n+\beta) - \psi(\beta) \end{aligned} $$

Tests

Test is written follow the guide.

Side Effects

No.

Release notes

beta_neg_binomial_lpmf is available if merged. Allows modeling heavy-tail count data.

Checklist

• Copyright holder: Zhi Ling

  The copyright holder is typically you or your assignee, such as a university or company. By submitting this pull request, the copyright holder is agreeing to the license the submitted work under the following licenses:
  - Code: BSD 3-clause (https://opensource.org/licenses/BSD-3-Clause)
  - Documentation: CC-BY 4.0 (https://creativecommons.org/licenses/by/4.0/)

• the basic tests are passing

  • unit tests pass (to run, use: ./runTests.py test/unit)
  • header checks pass, (make test-headers)
  • dependencies checks pass, (make test-math-dependencies)
  • docs build, (make doxygen)
  • code passes the built in C++ standards checks (make cpplint)

• the code is written in idiomatic C++ and changes are documented in the doxygen

• the new changes are tested

[stan-buildbot]

---

Name	Old Result	New Result	Ratio	Performance change( 1 - new / old )
arma/arma.stan	0.32	0.38	0.84	-18.63% slower
low_dim_corr_gauss/low_dim_corr_gauss.stan	0.01	0.01	0.98	-1.58% slower
gp_regr/gen_gp_data.stan	0.03	0.03	1.04	4.28% faster
gp_regr/gp_regr.stan	0.1	0.09	1.01	0.79% faster
sir/sir.stan	69.45	77.14	0.9	-11.07% slower
irt_2pl/irt_2pl.stan	4.18	3.96	1.05	5.13% faster
eight_schools/eight_schools.stan	0.06	0.05	1.07	6.9% faster
pkpd/sim_one_comp_mm_elim_abs.stan	0.25	0.24	1.06	5.5% faster
pkpd/one_comp_mm_elim_abs.stan	19.49	18.62	1.05	4.47% faster
garch/garch.stan	0.43	0.41	1.06	5.33% faster
low_dim_gauss_mix/low_dim_gauss_mix.stan	2.71	2.59	1.04	4.24% faster
arK/arK.stan	1.79	1.71	1.05	4.36% faster
gp_pois_regr/gp_pois_regr.stan	2.81	2.68	1.05	4.51% faster
low_dim_gauss_mix_collapse/low_dim_gauss_mix_collapse.stan	8.77	8.54	1.03	2.61% faster
performance.compilation	182.01	179.89	1.01	1.16% faster
Mean result: 1.0163824566291206

---

Jenkins Console Log
Blue Ocean
Commit hash: 2c8bf6d3d4fe497e1d8ff310ce5e5f26b1925faa

---

Machine information

No LSB modules are available. Distributor ID: Ubuntu Description: Ubuntu 20.04.3 LTS Release: 20.04 Codename: focal

CPU:
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Byte Order: Little Endian
Address sizes: 46 bits physical, 48 bits virtual
CPU(s): 80
On-line CPU(s) list: 0-79
Thread(s) per core: 2
Core(s) per socket: 20
Socket(s): 2
NUMA node(s): 2
Vendor ID: GenuineIntel
CPU family: 6
Model: 85
Model name: Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz
Stepping: 4
CPU MHz: 2400.000
CPU max MHz: 3700.0000
CPU min MHz: 1000.0000
BogoMIPS: 4800.00
Virtualization: VT-x
L1d cache: 1.3 MiB
L1i cache: 1.3 MiB
L2 cache: 40 MiB
L3 cache: 55 MiB
NUMA node0 CPU(s): 0,2,4,6,8,10,12,14,16,18,20,22,24,26,28,30,32,34,36,38,40,42,44,46,48,50,52,54,56,58,60,62,64,66,68,70,72,74,76,78
NUMA node1 CPU(s): 1,3,5,7,9,11,13,15,17,19,21,23,25,27,29,31,33,35,37,39,41,43,45,47,49,51,53,55,57,59,61,63,65,67,69,71,73,75,77,79
Vulnerability Gather data sampling: Mitigation; Microcode
Vulnerability Itlb multihit: KVM: Mitigation: VMX disabled
Vulnerability L1tf: Mitigation; PTE Inversion; VMX conditional cache flushes, SMT vulnerable
Vulnerability Mds: Mitigation; Clear CPU buffers; SMT vulnerable
Vulnerability Meltdown: Mitigation; PTI
Vulnerability Mmio stale data: Mitigation; Clear CPU buffers; SMT vulnerable
Vulnerability Reg file data sampling: Not affected
Vulnerability Retbleed: Mitigation; IBRS
Vulnerability Spec rstack overflow: Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2: Mitigation; IBRS; IBPB conditional; STIBP conditional; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected
Vulnerability Srbds: Not affected
Vulnerability Tsx async abort: Mitigation; Clear CPU buffers; SMT vulnerable
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb cat_l3 cdp_l3 invpcid_single pti intel_ppin ssbd mba ibrs ibpb stibp tpr_shadow vnmi flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm cqm mpx rdt_a avx512f avx512dq rdseed adx smap clflushopt clwb intel_pt avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local dtherm ida arat pln pts hwp hwp_act_window hwp_epp hwp_pkg_req pku ospke md_clear flush_l1d arch_capabilities

G++:
g++ (Ubuntu 9.4.0-1ubuntu1~20.04) 9.4.0
Copyright (C) 2019 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

Clang:
clang version 10.0.0-4ubuntu1
Target: x86_64-pc-linux-gnu
Thread model: posix
InstalledDir: /usr/bin

[WardBrian]

@lingium not a review itself, but wanted to chime in to say super nice work so far! It's no small feat to have a PR pass CI on the first try; I hardly do myself even now.

Nice meeting you at StanCon, and glad you're contributing to Stan!

[lingium]

@lingium not a review itself, but wanted to chime in to say super nice work so far! It's no small feat to have a PR pass CI on the first try; I hardly do myself even now.

Nice meeting you at StanCon, and glad you're contributing to Stan!

Hey @WardBrian, thanks for your support! It was a pleasure meeting you at StanCon as well. I'm glad to be contributing to Stan :)

[lingium]

Seems like I forgot to pull pre-commit hook's commit to local

[lingium]

May have to require your approval to one of the workflows to continue

[stan-buildbot]

---

Name	Old Result	New Result	Ratio	Performance change( 1 - new / old )
arma/arma.stan	0.35	0.34	1.01	0.94% faster
low_dim_corr_gauss/low_dim_corr_gauss.stan	0.01	0.01	0.91	-9.65% slower
gp_regr/gen_gp_data.stan	0.03	0.03	0.98	-2.42% slower
gp_regr/gp_regr.stan	0.1	0.1	1.01	1.21% faster
sir/sir.stan	74.33	74.3	1.0	0.04% faster
irt_2pl/irt_2pl.stan	4.77	4.26	1.12	10.68% faster
eight_schools/eight_schools.stan	0.06	0.06	1.06	5.28% faster
pkpd/sim_one_comp_mm_elim_abs.stan	0.27	0.26	1.04	3.47% faster
pkpd/one_comp_mm_elim_abs.stan	20.77	20.32	1.02	2.15% faster
garch/garch.stan	0.5	0.46	1.08	7.38% faster
low_dim_gauss_mix/low_dim_gauss_mix.stan	2.95	3.12	0.95	-5.53% slower
arK/arK.stan	1.92	1.8	1.07	6.15% faster
gp_pois_regr/gp_pois_regr.stan	3.02	2.92	1.04	3.43% faster
low_dim_gauss_mix_collapse/low_dim_gauss_mix_collapse.stan	10.12	8.93	1.13	11.78% faster
performance.compilation	199.87	201.22	0.99	-0.67% slower
Mean result: 1.026584005369048

---

Jenkins Console Log
Blue Ocean
Commit hash: 2c8bf6d3d4fe497e1d8ff310ce5e5f26b1925faa

---

Machine information

No LSB modules are available. Distributor ID: Ubuntu Description: Ubuntu 20.04.3 LTS Release: 20.04 Codename: focal

CPU:
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Byte Order: Little Endian
Address sizes: 46 bits physical, 48 bits virtual
CPU(s): 80
On-line CPU(s) list: 0-79
Thread(s) per core: 2
Core(s) per socket: 20
Socket(s): 2
NUMA node(s): 2
Vendor ID: GenuineIntel
CPU family: 6
Model: 85
Model name: Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz
Stepping: 4
CPU MHz: 3350.398
CPU max MHz: 3700.0000
CPU min MHz: 1000.0000
BogoMIPS: 4800.00
Virtualization: VT-x
L1d cache: 1.3 MiB
L1i cache: 1.3 MiB
L2 cache: 40 MiB
L3 cache: 55 MiB
NUMA node0 CPU(s): 0,2,4,6,8,10,12,14,16,18,20,22,24,26,28,30,32,34,36,38,40,42,44,46,48,50,52,54,56,58,60,62,64,66,68,70,72,74,76,78
NUMA node1 CPU(s): 1,3,5,7,9,11,13,15,17,19,21,23,25,27,29,31,33,35,37,39,41,43,45,47,49,51,53,55,57,59,61,63,65,67,69,71,73,75,77,79
Vulnerability Gather data sampling: Mitigation; Microcode
Vulnerability Itlb multihit: KVM: Mitigation: VMX disabled
Vulnerability L1tf: Mitigation; PTE Inversion; VMX conditional cache flushes, SMT vulnerable
Vulnerability Mds: Mitigation; Clear CPU buffers; SMT vulnerable
Vulnerability Meltdown: Mitigation; PTI
Vulnerability Mmio stale data: Mitigation; Clear CPU buffers; SMT vulnerable
Vulnerability Reg file data sampling: Not affected
Vulnerability Retbleed: Mitigation; IBRS
Vulnerability Spec rstack overflow: Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2: Mitigation; IBRS; IBPB conditional; STIBP conditional; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected
Vulnerability Srbds: Not affected
Vulnerability Tsx async abort: Mitigation; Clear CPU buffers; SMT vulnerable
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb cat_l3 cdp_l3 invpcid_single pti intel_ppin ssbd mba ibrs ibpb stibp tpr_shadow vnmi flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm cqm mpx rdt_a avx512f avx512dq rdseed adx smap clflushopt clwb intel_pt avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local dtherm ida arat pln pts hwp hwp_act_window hwp_epp hwp_pkg_req pku ospke md_clear flush_l1d arch_capabilities

G++:
g++ (Ubuntu 9.4.0-1ubuntu1~20.04) 9.4.0
Copyright (C) 2019 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

Clang:
clang version 10.0.0-4ubuntu1
Target: x86_64-pc-linux-gnu
Thread model: posix
InstalledDir: /usr/bin

[stan-buildbot]

---

Name	Old Result	New Result	Ratio	Performance change( 1 - new / old )
arma/arma.stan	0.35	0.4	0.88	-14.06% slower
low_dim_corr_gauss/low_dim_corr_gauss.stan	0.01	0.01	0.91	-9.7% slower
gp_regr/gen_gp_data.stan	0.03	0.03	0.97	-3.13% slower
gp_regr/gp_regr.stan	0.12	0.1	1.15	13.15% faster
sir/sir.stan	76.11	76.9	0.99	-1.03% slower
irt_2pl/irt_2pl.stan	5.11	4.78	1.07	6.52% faster
eight_schools/eight_schools.stan	0.06	0.06	1.06	5.3% faster
pkpd/sim_one_comp_mm_elim_abs.stan	0.28	0.26	1.05	4.52% faster
pkpd/one_comp_mm_elim_abs.stan	21.97	20.38	1.08	7.22% faster
garch/garch.stan	0.51	0.47	1.08	7.42% faster
low_dim_gauss_mix/low_dim_gauss_mix.stan	3.04	2.88	1.06	5.25% faster
arK/arK.stan	2.06	1.9	1.09	8.0% faster
gp_pois_regr/gp_pois_regr.stan	3.38	3.12	1.08	7.55% faster
low_dim_gauss_mix_collapse/low_dim_gauss_mix_collapse.stan	10.36	9.4	1.1	9.26% faster
performance.compilation	202.66	199.59	1.02	1.51% faster
Mean result: 1.0381099484857623

---

Jenkins Console Log
Blue Ocean
Commit hash: 2c8bf6d3d4fe497e1d8ff310ce5e5f26b1925faa

---

Machine information

No LSB modules are available. Distributor ID: Ubuntu Description: Ubuntu 20.04.3 LTS Release: 20.04 Codename: focal

CPU:
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Byte Order: Little Endian
Address sizes: 46 bits physical, 48 bits virtual
CPU(s): 80
On-line CPU(s) list: 0-79
Thread(s) per core: 2
Core(s) per socket: 20
Socket(s): 2
NUMA node(s): 2
Vendor ID: GenuineIntel
CPU family: 6
Model: 85
Model name: Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz
Stepping: 4
CPU MHz: 2400.000
CPU max MHz: 3700.0000
CPU min MHz: 1000.0000
BogoMIPS: 4800.00
Virtualization: VT-x
L1d cache: 1.3 MiB
L1i cache: 1.3 MiB
L2 cache: 40 MiB
L3 cache: 55 MiB
NUMA node0 CPU(s): 0,2,4,6,8,10,12,14,16,18,20,22,24,26,28,30,32,34,36,38,40,42,44,46,48,50,52,54,56,58,60,62,64,66,68,70,72,74,76,78
NUMA node1 CPU(s): 1,3,5,7,9,11,13,15,17,19,21,23,25,27,29,31,33,35,37,39,41,43,45,47,49,51,53,55,57,59,61,63,65,67,69,71,73,75,77,79
Vulnerability Gather data sampling: Mitigation; Microcode
Vulnerability Itlb multihit: KVM: Mitigation: VMX disabled
Vulnerability L1tf: Mitigation; PTE Inversion; VMX conditional cache flushes, SMT vulnerable
Vulnerability Mds: Mitigation; Clear CPU buffers; SMT vulnerable
Vulnerability Meltdown: Mitigation; PTI
Vulnerability Mmio stale data: Mitigation; Clear CPU buffers; SMT vulnerable
Vulnerability Reg file data sampling: Not affected
Vulnerability Retbleed: Mitigation; IBRS
Vulnerability Spec rstack overflow: Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2: Mitigation; IBRS; IBPB conditional; STIBP conditional; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected
Vulnerability Srbds: Not affected
Vulnerability Tsx async abort: Mitigation; Clear CPU buffers; SMT vulnerable
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb cat_l3 cdp_l3 invpcid_single pti intel_ppin ssbd mba ibrs ibpb stibp tpr_shadow vnmi flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm cqm mpx rdt_a avx512f avx512dq rdseed adx smap clflushopt clwb intel_pt avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local dtherm ida arat pln pts hwp hwp_act_window hwp_epp hwp_pkg_req pku ospke md_clear flush_l1d arch_capabilities

G++:
g++ (Ubuntu 9.4.0-1ubuntu1~20.04) 9.4.0
Copyright (C) 2019 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

Clang:
clang version 10.0.0-4ubuntu1
Target: x86_64-pc-linux-gnu
Thread model: posix
InstalledDir: /usr/bin

[SteveBronder]

Looks good to me! Will github let you press the merge button? If not I can, though we use to be able to let contributors click the button themselves which imo feels nice

[lingium]

Thanks for the review @SteveBronder! From my side it says merging is blocked, so please go ahead from your end

[SteveBronder]

@lingium check your email. You should have received an invite to the stan org and with that I believe you will be able to click the merge button :)

[lingium]

Hi @SteveBronder, got the invite. I was tagged as a collaborator now but don't seem to have merge pr permission

[WardBrian]

@lingium maybe now?

[lingium]

Yes, it's working now!

[WardBrian]

@lingium are you also interested in adding this to the compiler yourself, or would you prefer another developer do that part? It's easy enough for me to do if you want to provide a list of signatures you'd expect support for

[lingium]

@WardBrian could you help me to add this time? Cuz it's super easy for you. Here's list of signatures

real beta_neg_binomial_lpmf(int n, real r, real alpha, real beta);
real beta_neg_binomial_lpmf(int n, vector r, real alpha, real beta);
real beta_neg_binomial_lpmf(int n, real r, vector alpha, real beta);
real beta_neg_binomial_lpmf(int n, real r, real alpha, vector beta);
real beta_neg_binomial_lpmf(int n, vector r, vector alpha, real beta);
real beta_neg_binomial_lpmf(int n, vector r, real alpha, vector beta);
real beta_neg_binomial_lpmf(int n, real r, vector alpha, vector beta);
real beta_neg_binomial_lpmf(int n, vector r, vector alpha, vector beta);
real beta_neg_binomial_lpmf(array[] int n, real r, real alpha, real beta);
real beta_neg_binomial_lpmf(array[] int n, vector r, real alpha, real beta);
real beta_neg_binomial_lpmf(array[] int n, real r, vector alpha, real beta);
real beta_neg_binomial_lpmf(array[] int n, real r, real alpha, vector beta);
real beta_neg_binomial_lpmf(array[] int n, vector r, vector alpha, real beta);
real beta_neg_binomial_lpmf(array[] int n, vector r, real alpha, vector beta);
real beta_neg_binomial_lpmf(array[] int n, real r, vector alpha, vector beta);
real beta_neg_binomial_lpmf(array[] int n, vector r, vector alpha, vector beta);

Another thing is, we would prefer small PRs, right? If so for the cdf, lcdf, lccdf, and rng, I will submit in different PRs.

[WardBrian]

I’ll take a look at that today!

And yes, it is easiest if they are each in their own PR. If I remember correctly from Stan Con, isn’t there also an alternative parameterization you are interested in?

[lingium]

Thanks Brian!

Yes, there is an alternative parameterization. It becomes neg_binomial_2 when one of the parameters is equal to 0. Except for that, I'm planning Waring and Yule distribution, both are special cases of beta negative binomial.