P. 1
Fuzzy Time Series Forecasting with Average Based Interval Thesis

# Fuzzy Time Series Forecasting with Average Based Interval Thesis

Ratings: (0)|Views: 1,214 |Likes:
Forecasting systems with fuzzy time series capturing the
pattern of past data and then use it to project future data. The
process also does not require a complex learning system as it exists
on genetic algorithms and neural networks, so that make the system
is easy to develop. In the prediction using fuzzy time series, the
length of the interval has been determined at the beginning of the
calculation process. While determining the interval length is very
influential in the formation of fuzzy relationships also will have an
impact on the prediction of the outcome differences. Therefore, the
formation of the fuzzy relationship must be precise and it requires
the determination of an appropriate interval length. One method that
can be used to determine the effective length of the interval is an
average based method. In this paper, the authors implement the fuzzy
time series to forecast the monthly sales data, as for the data used for
testing is derived from census statistical data website provider. And
from the results of tests conducted that data forecasting using
Average-based Fuzzy Time Series with a determination based on the
average interval has a higher accuracy compared with Fuzzy Time
Series Standards, with the difference in average 52.39% more
accurate if the error is calculated using AFER and the difference in
average 70.90% more accurate if the error is calculated using the
MSE.
Forecasting systems with fuzzy time series capturing the
pattern of past data and then use it to project future data. The
process also does not require a complex learning system as it exists
on genetic algorithms and neural networks, so that make the system
is easy to develop. In the prediction using fuzzy time series, the
length of the interval has been determined at the beginning of the
calculation process. While determining the interval length is very
influential in the formation of fuzzy relationships also will have an
impact on the prediction of the outcome differences. Therefore, the
formation of the fuzzy relationship must be precise and it requires
the determination of an appropriate interval length. One method that
can be used to determine the effective length of the interval is an
average based method. In this paper, the authors implement the fuzzy
time series to forecast the monthly sales data, as for the data used for
testing is derived from census statistical data website provider. And
from the results of tests conducted that data forecasting using
Average-based Fuzzy Time Series with a determination based on the
average interval has a higher accuracy compared with Fuzzy Time
Series Standards, with the difference in average 52.39% more
accurate if the error is calculated using AFER and the difference in
average 70.90% more accurate if the error is calculated using the
MSE.

Published by: Syauqi Haris on Oct 22, 2011

### Availability:

Read on Scribd mobile: iPhone, iPad and Android.
See more
See less

06/13/2013

text

JI\HCICF_M^J IC_GOC @STTZ _JIC ^C[JC^ OCFEMF\CFCF_SMF JF_C[PMH KC[KM^J^ [M_M)[M_M SF_SB\C[MIMHMF OM_M \CFDSMHMF KSHMFMF

^B[J\^J
ghca4
I ^ZMS]J AM[J^<>=<:0<<?1):0\[GE[MI ^_SOJ JHIS BGI\S_C[DS[S^MF IM_CIM_JBM@MBSH_M^ IM_CIM_JBM OMF JHIS \CFEC_MASMF MHMISFJPC[^J_M^ K[MUJDMZMIMHMFE;<=<

jj

JI\HCICF_M^J IC_GOC @STTZ _JIC ^C[JC^ OCFEMF\CFCF_SMF JF_C[PMH KC[KM^J^ [M_M)[M_M SF_SB\C[MIMHMF OM_M \CFDSMHMF KSHMFMF

^B[J\^J
^ckmemj ~mhma ~mts ~zm{mt sftsb ici|c{ghca echm{^m{dmfm ^mjf~ omhmi kjomfe Jhis Bgi|stc{ghca4
I ^ZMS]J AM[J^<>=<:0<<?1):0\[GE[MI ^_SOJ JHIS BGI\S_C[DS[S^MF IM_CIM_JBM@MBSH_M^ IM_CIM_JBM OMF JHIS \CFEC_MASMF MHMISFJPC[^J_M^ K[MUJDMZMIMHMFE;<=<

## Activity (28)

You've already reviewed this. Edit your review.