Support Vector Regression Modelling for Rainfall Prediction in Dry Season Based on Southern Oscillation Index and NIN03.4
ICACSIS 2013
ISBN: 978-979-1421-19-5zyxwvutsrqponmlkjih
Support Vector Regression M odelling for Rainfall
Prediction in Dry Season Based on
Southern O scillation Index and NIN03.4
Gita Adhani'. Agus Buono', Akhmad Faqih2zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPON
IzyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
Department of Computer Science. -'Department of Geophysics and Meteorology
Faculty of Mathematics (Ind Natural Sciences. Bogor Agricultural University
Email : [email protected]@[email protected]
on elimate condition and weather. Climate and weather
factors to suceesses
agricultural
and
are crucial
plantation. Knowledge of climate patterns and weather
can help in making deeisions eropping patterns and
plant varieties appropriate in difTerent areas.
Various clirnate disasters in lndonesia are mostly
related to the El Nino Southern Oscillation
(EN50)
phenornenon. Climate variability, especially rainfall, is
strongly related to jhis phenornenon. Generally,
El
Nino impact on rainfall deereasing
or even drought.
otherwise
La Nina influenees on rainfall increasing
which can cause flooding [I].
La Nina causes curnrnulation
of air rnass that
contains
alot
of water vapor in the lndonesia
atrnosphere
thus potency of rain clouds fonning
enhanccs. As a result, although the middle of 20 I 0 dry
season, it still could be raining in rnany regions with
low up to high intensity [2].
El Nino phenornenon
gives more serious impact
than La Nina. El Nino causes rainfall in most area in
lndonesia
rcduced. This rainfall decreasing rate IS
really dependent on intensity and El Nino duration. El
111
Nino is noted onee caused long-term
drought
lndonesia. Rainfall information cluring dry season
is
greatly needed in agricultural and plantations. Rainfall
I. INTRODUCTION
forecasting
during
dry season
can be used as
information
for
larmers
to
mitigate
any cases that can
Clirnate is one of natural ecosystem cornponents that
be
happcned
like
preproduction
drought
that lead to
has a rnajor influence on the various sectors of human
crop
failure.
life. Indone ia as an agricultural country is dependent
This research porpose to forecast rainfall during dr)'
season by took case study in Indramayu region using
Manuscript
rcccivcd June 28. 2013. This wor],zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
\-a~ supponcd 111
Support
Vector
Regression
(SVR)
and related
part h~ the Computer Science Department
of Bo~('r :\~ricultllral
v ariables uscd are Southern Oscillation
Index (501)
l inivcrvrtv
(i3:\lJ).
Center
for Climatc
Risk and Opportunuy
and sea surface ternperature
(SST) at NINO 3.4
Management
ln Southcast Asia and Pacific (CCRO"I·Sh\J».
Bogor Agncultural
Univer sir, (13:\11), United States :\genc\
for
region. SVR is Support Vector Machine (SVM) is
Intemational
Development
(US/\ID).
used for regression case.
Bogor
Gita Adham IS WIth the Computer SCience Department.
Regrcssion
is one of cornmon season prediction
Agricultural
I inivcrsitv.
PO
BOX
I A6RO
11'f)O:"FSIA
methods. Support Vector Machine (SVM) is used to
I corrcsponduig
author to prol Ide phone "6~8~J(,60 I 59R-l. c-rnail
adharu gita II gmarl.cornj
solvc barriers in statistical rcgression analysis. Lincar
Ag'" I\1I"no IS with the Com ril ter Science Department.
B()~nr
regression based on several assumptions thus there can
Agncultural
1'ruvcrsuv.
PO nox 166RO INIX)~I·.SIA
(e-mail
not always
suitable
to the exisung
data set
pudcsha 1/ grnarl com)
charactcristics.
Formcrs research that applied SVR
Akhmad
laqrh
IS wrth the (ieophyslcs
and ~letcor('i
ISBN: 978-979-1421-19-5zyxwvutsrqponmlkjih
Support Vector Regression M odelling for Rainfall
Prediction in Dry Season Based on
Southern O scillation Index and NIN03.4
Gita Adhani'. Agus Buono', Akhmad Faqih2zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPON
IzyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
Department of Computer Science. -'Department of Geophysics and Meteorology
Faculty of Mathematics (Ind Natural Sciences. Bogor Agricultural University
Email : [email protected]@[email protected]
on elimate condition and weather. Climate and weather
factors to suceesses
agricultural
and
are crucial
plantation. Knowledge of climate patterns and weather
can help in making deeisions eropping patterns and
plant varieties appropriate in difTerent areas.
Various clirnate disasters in lndonesia are mostly
related to the El Nino Southern Oscillation
(EN50)
phenornenon. Climate variability, especially rainfall, is
strongly related to jhis phenornenon. Generally,
El
Nino impact on rainfall deereasing
or even drought.
otherwise
La Nina influenees on rainfall increasing
which can cause flooding [I].
La Nina causes curnrnulation
of air rnass that
contains
alot
of water vapor in the lndonesia
atrnosphere
thus potency of rain clouds fonning
enhanccs. As a result, although the middle of 20 I 0 dry
season, it still could be raining in rnany regions with
low up to high intensity [2].
El Nino phenornenon
gives more serious impact
than La Nina. El Nino causes rainfall in most area in
lndonesia
rcduced. This rainfall decreasing rate IS
really dependent on intensity and El Nino duration. El
111
Nino is noted onee caused long-term
drought
lndonesia. Rainfall information cluring dry season
is
greatly needed in agricultural and plantations. Rainfall
I. INTRODUCTION
forecasting
during
dry season
can be used as
information
for
larmers
to
mitigate
any cases that can
Clirnate is one of natural ecosystem cornponents that
be
happcned
like
preproduction
drought
that lead to
has a rnajor influence on the various sectors of human
crop
failure.
life. Indone ia as an agricultural country is dependent
This research porpose to forecast rainfall during dr)'
season by took case study in Indramayu region using
Manuscript
rcccivcd June 28. 2013. This wor],zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
\-a~ supponcd 111
Support
Vector
Regression
(SVR)
and related
part h~ the Computer Science Department
of Bo~('r :\~ricultllral
v ariables uscd are Southern Oscillation
Index (501)
l inivcrvrtv
(i3:\lJ).
Center
for Climatc
Risk and Opportunuy
and sea surface ternperature
(SST) at NINO 3.4
Management
ln Southcast Asia and Pacific (CCRO"I·Sh\J».
Bogor Agncultural
Univer sir, (13:\11), United States :\genc\
for
region. SVR is Support Vector Machine (SVM) is
Intemational
Development
(US/\ID).
used for regression case.
Bogor
Gita Adham IS WIth the Computer SCience Department.
Regrcssion
is one of cornmon season prediction
Agricultural
I inivcrsitv.
PO
BOX
I A6RO
11'f)O:"FSIA
methods. Support Vector Machine (SVM) is used to
I corrcsponduig
author to prol Ide phone "6~8~J(,60 I 59R-l. c-rnail
adharu gita II gmarl.cornj
solvc barriers in statistical rcgression analysis. Lincar
Ag'" I\1I"no IS with the Com ril ter Science Department.
B()~nr
regression based on several assumptions thus there can
Agncultural
1'ruvcrsuv.
PO nox 166RO INIX)~I·.SIA
not always
suitable
to the exisung
data set
pudcsha 1/ grnarl com)
charactcristics.
Formcrs research that applied SVR
Akhmad
laqrh
IS wrth the (ieophyslcs
and ~letcor('i