[
https://issues.apache.org/jira/browse/IGNITE-7438?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Anton Dmitriev updated IGNITE-7438:
-----------------------------------
Description:
This task consists of two parts:
* Implementation of the +LSQR iterative solver+ of systems of linear equations.
* Implementation of the +LSQR-based linear regression trainer+.
Apache Ignite LSQR iterative solver is based on [SciPy reference
implementation|http://example.com/], but it's distributed and can:
* Efficiently work in cases when a data is distributed across a cluster.
* Utilize all CPU resources by processing different parts of data on different
cores.
These advantages are achieved as result of changing [Golub-Kahan-Lanczos
Bidiagonalization
Procedure|http://www.netlib.org/utk/people/JackDongarra/etemplates/node198.html]
procedure which is a core of LSQR algorithm and utilizing features of
[Partition Based Dataset
implementation|https://issues.apache.org/jira/browse/IGNITE-7437].
LSQR-based linear regression trainer is a trainer that uses LSQR solver to
solve system of linear equations which represents linear regression problem.
was:
This task consists of two parts:
* Implementation of the +LSQR iterative solver+ of systems of linear equations.
* Implementation of the +LSQR-based linear regression trainer+.
Apache Ignite LSQR iterative solver is based on [SciPy reference
implementation|http://example.com/], but it's distributed and can:
* Efficiently work in cases when a data is distributed across a cluster.
* Utilize all CPU resources by processing different parts of data on different
cores.
These advantages are achieved as result of changing [Golub-Kahan-Lanczos
Bidiagonalization
Procedure|http://www.netlib.org/utk/people/JackDongarra/etemplates/node198.html]
procedure which is a core of LSQR algorithm and utilizing features of
Partition Based Dataset implementation.
LSQR-based linear regression trainer is a trainer that uses LSQR solver to
solve system of linear equations which represents linear regression problem.
> LSQR: Sparse Equations and Least Squares for Lin Regression
> -----------------------------------------------------------
>
> Key: IGNITE-7438
> URL: https://issues.apache.org/jira/browse/IGNITE-7438
> Project: Ignite
> Issue Type: New Feature
> Components: ml
> Reporter: Yury Babak
> Assignee: Anton Dmitriev
> Priority: Major
> Fix For: 2.5
>
>
> This task consists of two parts:
> * Implementation of the +LSQR iterative solver+ of systems of linear
> equations.
> * Implementation of the +LSQR-based linear regression trainer+.
>
> Apache Ignite LSQR iterative solver is based on [SciPy reference
> implementation|http://example.com/], but it's distributed and can:
> * Efficiently work in cases when a data is distributed across a cluster.
> * Utilize all CPU resources by processing different parts of data on
> different cores.
> These advantages are achieved as result of changing [Golub-Kahan-Lanczos
> Bidiagonalization
> Procedure|http://www.netlib.org/utk/people/JackDongarra/etemplates/node198.html]
> procedure which is a core of LSQR algorithm and utilizing features of
> [Partition Based Dataset
> implementation|https://issues.apache.org/jira/browse/IGNITE-7437].
>
> LSQR-based linear regression trainer is a trainer that uses LSQR solver to
> solve system of linear equations which represents linear regression problem.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
