Skip to main content
Home
Browse All
Log in

Favorites

Help

English
English
EngishPirate
한국어
Search
Advanced Search
Find results with:
error div
Add another field
Search by date
Search by date:
from
after
before
on
from:
to
to:
Searching collections:
CCSU Theses and Dissertations
Add or remove collections
Home
CCSU Theses & Dissertations
Use of parallel computing to fit OLS Regression models using SAS
Reference URL
Share
Add tags
Comment
Rate
Save to favorites
Remove from favorites
To link to this object, paste this link in email, IM or document
To embed this object, paste this HTML in website
Use of parallel computing to fit OLS Regression models using SAS
View Description
Download
small (250x250 max)
medium (500x500 max)
Large
Extra Large
large ( > 500x500)
Full Resolution
Print
This item is restricted to only allow viewing of the metadata.
Description
Identifier
Thesis
2499
Author
Souza, Marcos J.M. de (Marcos José), 1976
Title
Use
of
parallel
computing
to
fit
OLS
Regression
models
using
SAS
Publisher
Central Connecticut State University
Date of Publication
2015
Resource Type
Master's Thesis
Abstract
The
regression
analysis
technique
allows
developing
a
model
to
forecast
the
values
of a
numerical
variable
using
as
input
the
value
of
other
variables
. In the
regression
analysis
, the
variable
under
study
is
called
dependent
variable
, and the
variables
used
as
input
are
called
independent
variables
.
Besides
forecasting
the
values
for the
dependent
variable
,
regression
analysis
also
allows
to
identify
the
type
of
relationship
(linear
,
quadratic
,
etc.)
that
exists
between
the
dependent
variable
and in the
independent
variables
,
quantify
the
effect
that
changes
in the
independent
variables
will
cause
in the
dependent
variables
, and also
identify
outliers
.
Diverse
organizations
use
regression
analysis
in their
daytoday
businesses
, for
many
different
purposes
.
Between
many
other
examples
,
governments
utilize
regression
analysis
to
identify
fraud
on
taxes
declarations
,
pharmaceutical
industries
utilize
regression
analysis
in the
clinical
trials
,
marketing
companies
utilize
regression
to
target
population
, and
banks
utilize
regression
to
forecast
delinquency
.
Pritchard
and
Kahn
[14]
explain
how
regression
was
used
to
target
fraudulent
taxes
declarations
in the
UK
.
McCausland
and
Sedgley
[15]
present
a
study
using
regression
analysis
to
determine
lost
revenue
to a
division
of a
small
public
relations
firm
. The
physical
universe
is
large
as
is
the
data
universe
–
by
2020
the
latter
will
contain
, as
nearly
as
many
digital
bits
as there are
stars
in the
universe
.
It’s
doubling
in
size
every
two
years
, and by
2020
the
digital
universe
–
the
data
we
create
and
copy
annually
–
will
reach
44
zettabytes
, or
44
trillion
gigabytes
[13]
. The
architecture
used
in the
modern
computer
chips
is
reaching
physical
limits
,
mainly
because
of
cooling
requirements
. In the
past
few
years
,
CPUs
have not
improved
much
in
clock
speed
, but
more
cores
have been
added
to
chips
. As
it
won’t
be
possible
to
compute
with
faster
CPUs
(not
using
the
actual
technology)
,
it’s
necessary
to
break
processes
in
multiple
parts
, to
achieve
a
reasonable
total
computational
run
time
,
while
data
grows
and
business
requires
additional
analysis
. This
research
presents
a
method
of
parallelizing
the
computation
of the
regression
analysis
across
multiple
CPUs
and
computers
, to
allow
reducing
total
runtime
and
processing
very
large
databases
. Not
every
problem
that
is
parallelized
runs
faster
than its
nonparallel
version
.
Although
, as
computers
become
such
a
commodity
asset
,
building
large
computing
clusters
and
distributing
the
computation
to
fit
models
across
all
of them
is
a
feature
that
analysts
will have at their
hands
. This
research
won’t
demonstrate
the
turning
point
,
where
parallelizing
becomes
faster
than
running
a
nonparallel
version
, but
it
will
demonstrate
that the
algorithm
developed
will
scale
as
more
computational
resources
are
added
, by
processing
the
same
amount
of
data
faster
.
Notes
"
Submitted
in
Partial
Fulfillment
of the
Requirements
for the
Degree
of
Master
of
Science
in
Data
Mining
,
Department
of
Mathematical
Sciences.
";
Thesis
advisor
:
Roger
Bilisoly.
;
M.S.,Central
Connecticut
State
University,,2015.
;
Includes
bibliographical
references
(leaves
4243)
.
Subject
SAS (Computer file)
Parallel computers.
Regression analysis.
Department
Department of Mathematical Sciences
Advisor
Bilisoly, Roger, 1963
Type
Text
Digital Format
application/pdf
Software
System requirements: PC and World Wide Web browser.
Language
eng
OCLC number
940975653
Rating
Tags
Add tags
for Use of parallel computing to fit OLS Regression models using SAS
View as list

View as tag cloud

report abuse
Comments
Post a Comment
for
Use of parallel computing to fit OLS Regression models using SAS
Your rating was saved.
you wish to report:
Your comment:
Your Name:
...
Back to top
Select the collections to add or remove from your search
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
X
Y
Z
Select All Collections
C
CCSU Student Publications
CCSU Theses and Dissertations
G
GLBTQ Archives
M
Modern Language Oral Histories
O
O'Neill Archives Oral Histories
P
Polish American Pamphlets
Polish Posters
T
Treasures from the Special Collections
V
Veterans History Project
500
You have selected:
1
OK
Cancel