Date: Sun, 3 Jan 2021 19:57:28 +0000 (UTC) From: Sunpoet Po-Chuan Hsieh <sunpoet@FreeBSD.org> To: ports-committers@freebsd.org, svn-ports-all@freebsd.org, svn-ports-head@freebsd.org Subject: svn commit: r560045 - in head/math: . py-spglm Message-ID: <202101031957.103JvSWN086570@repo.freebsd.org>
next in thread | raw e-mail | index | archive | help
Author: sunpoet Date: Sun Jan 3 19:57:28 2021 New Revision: 560045 URL: https://svnweb.freebsd.org/changeset/ports/560045 Log: Add py-spglm 1.0.8 This module is an adaptation of a portion of GLM functionality from the Statsmodels package, this it has been simplified and customized for the purposes of serving as the base for several other PySAL modules, namely SpInt and GWR. Currently, it supports the estimation of Gaussian, Poisson, and Logistic regression using only iteratively weighted least squares estimation (IWLS). One of the large differences this module and the functions avaialble in the Statsmodels package is that the custom IWLS routine is fully sparse compatible, which was necesary for the very sparse design matrices that arise in constrained spatial interaction models. The somewhat limited functionality and computation of only a subset of GLM diagnostics also decreases the computational overhead. Another difference is that this module also supports the estimation of QuasiPoisson models. One caveat is that this custom IWLS routine currently generates estimates by directly solves the least squares normal equations rather than using a more robust method like the pseudo inverse. For more robust estimation of ill conditioned data and a fuller GLM framework we suggest using the original GLM functionality from Statsmodels. WWW: https://github.com/pysal/spglm Added: head/math/py-spglm/ head/math/py-spglm/Makefile (contents, props changed) head/math/py-spglm/distinfo (contents, props changed) head/math/py-spglm/pkg-descr (contents, props changed) Modified: head/math/Makefile Modified: head/math/Makefile ============================================================================== --- head/math/Makefile Sun Jan 3 19:57:22 2021 (r560044) +++ head/math/Makefile Sun Jan 3 19:57:28 2021 (r560045) @@ -843,6 +843,7 @@ SUBDIR += py-simhash SUBDIR += py-snuggs SUBDIR += py-spectral + SUBDIR += py-spglm SUBDIR += py-spot SUBDIR += py-ssm SUBDIR += py-statsmodels Added: head/math/py-spglm/Makefile ============================================================================== --- /dev/null 00:00:00 1970 (empty, because file is newly added) +++ head/math/py-spglm/Makefile Sun Jan 3 19:57:28 2021 (r560045) @@ -0,0 +1,25 @@ +# Created by: Po-Chuan Hsieh <sunpoet@FreeBSD.org> +# $FreeBSD$ + +PORTNAME= spglm +PORTVERSION= 1.0.8 +CATEGORIES= math python +MASTER_SITES= CHEESESHOP +PKGNAMEPREFIX= ${PYTHON_PKGNAMEPREFIX} + +MAINTAINER= sunpoet@FreeBSD.org +COMMENT= Sparse generalize linear models + +LICENSE= BSD3CLAUSE + +RUN_DEPENDS= ${PYTHON_PKGNAMEPREFIX}libpysal>=4.0.0:science/py-libpysal@${PY_FLAVOR} \ + ${PYTHON_PKGNAMEPREFIX}numpy>=1.3,1:math/py-numpy@${PY_FLAVOR} \ + ${PYTHON_PKGNAMEPREFIX}scipy>=0.11:science/py-scipy@${PY_FLAVOR} \ + ${PYTHON_PKGNAMEPREFIX}spreg>=1.0.4:math/py-spreg@${PY_FLAVOR} + +USES= python:3.6+ +USE_PYTHON= autoplist concurrent distutils + +NO_ARCH= yes + +.include <bsd.port.mk> Added: head/math/py-spglm/distinfo ============================================================================== --- /dev/null 00:00:00 1970 (empty, because file is newly added) +++ head/math/py-spglm/distinfo Sun Jan 3 19:57:28 2021 (r560045) @@ -0,0 +1,3 @@ +TIMESTAMP = 1609598745 +SHA256 (spglm-1.0.8.tar.gz) = df83b8f7caa41c8aebc4cc39179e40e8670783b06ee567b59bbbe818b773f300 +SIZE (spglm-1.0.8.tar.gz) = 37240 Added: head/math/py-spglm/pkg-descr ============================================================================== --- /dev/null 00:00:00 1970 (empty, because file is newly added) +++ head/math/py-spglm/pkg-descr Sun Jan 3 19:57:28 2021 (r560045) @@ -0,0 +1,18 @@ +This module is an adaptation of a portion of GLM functionality from the +Statsmodels package, this it has been simplified and customized for the purposes +of serving as the base for several other PySAL modules, namely SpInt and GWR. +Currently, it supports the estimation of Gaussian, Poisson, and Logistic +regression using only iteratively weighted least squares estimation (IWLS). One +of the large differences this module and the functions avaialble in the +Statsmodels package is that the custom IWLS routine is fully sparse compatible, +which was necesary for the very sparse design matrices that arise in constrained +spatial interaction models. The somewhat limited functionality and computation +of only a subset of GLM diagnostics also decreases the computational overhead. +Another difference is that this module also supports the estimation of +QuasiPoisson models. One caveat is that this custom IWLS routine currently +generates estimates by directly solves the least squares normal equations rather +than using a more robust method like the pseudo inverse. For more robust +estimation of ill conditioned data and a fuller GLM framework we suggest using +the original GLM functionality from Statsmodels. + +WWW: https://github.com/pysal/spglm
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?202101031957.103JvSWN086570>