Page 91 - Contributed Paper Session (CPS) - Volume 6

CPS1832 Nur Fazliana Rahim et al.
FER will surely be a great assist for the government in preparing for the
instability of the exchange rate as such. Furthermore, as it’s a major player in
forecasting of exchange rates, Bank Negara Malaysia (BNM) will also obtain
the benefit of this research. In the near future, the use of forecasting exchange
rates will surely be made better.
Based on the prior study, a new FTS model was proposed by Yu (2005)
during the creation of FTS model to modify the intervals length. As shown by
the outcome, appropriate fuzzy relationships can be improved the forecasting
values. Another study by Arumugam et al. (2013) was conducted in order to
predict Taiwan export trade using FTS model and ARIMA model. The final
results indicate that, FTS model beat ARIMA model with smallest average error
and its ability to forecast the Taiwan export trade. A viable practice to forecast
correctly and successfully future exchange were needed by the government
in dealing with foreign exchange rate as it can affect the country currency.
Even though lots of forecasting method were used to predict the foreign
exchange rate, still there are no way to show which are the most reliable
forecasting method. As stated by Applanaidu et al. (2011), the selection
methods should take into account several aspects, such as statistical data,
financing, level of accuracy and its significance. Nonetheless, these models are
quite expensive, need a great degree of expertise and numerous types of data
which may not always be obtainable.
Fuzzy rule-based systems (FRBS) deliver impulsive technique of reasoning
based on linguistic models (Dubois and Prade, 2001). Often reasoning based
on fuzzy models contribute an option to handle all kinds of unspecific data,
which presented the way people think and make decisions. (Rasmani and
Shen, 2006). In fuzzy forecasting, the determination of fuzzy rules is one of the
aspect to reflect on. The FTS rises the precision of the results made in
predicting situations that involve subjective, ambiguous and inaccurate
information by implementing these rules. So, in conjunction with this research,
Weighted Subsethood-Based Algorithm (WSBA) is suitable for generating
fuzzy rules for two reasons, which are easiness of the method and possible to
produce higher degree of accuracy indirectly minimize rules compared to
other methods (Chen and Tsai, 2008). The development of WSBA includes a
modification of the Subsethood-Based Algorithm (SBA) and the use of fuzzy
general rules and SBA values as weight.

2. Methodology
This section explained and discussed in detail four parts that have been
carried out in the research. The detailed are as follows.
A. Part 1: Compilation and Processing Data
Foreign Exchange Rate (FER) data is collected by referring to the to the
secondary data of monthly FER for five years’ data from year 2010 to 2015.

80 | I S I W S C 2 0 1 9