Information Technology – Thesis and Papers
- Applying Hadoop’s MapReduce Framework on Clustering the GPS Signals through Cloud Computing
By Wichian Premchaiswadi, Walisa Romsaiyud, Sarayut Intarasema / Graduate School of Information Technology, Siam University, Bangkok, Thailand. / Click here to view the report
1. Probabilistic Model of Nanometer MIFGMOSFET
Rawid Banchuin and Roungsan Chaisricharoen
The probabilistic model of the random variation in drain current of the nanometer multiple input floating-gate MOSFET (MIFGMOSFET) has been proposed in this research. The modeling process has taken the major physical level causes of random variations e.g. random dopant fluctuation and line edge roughness etc., into account. The proposed model have been found to be very accurate since it can fit the probabilistic distributions of normalized random drain current variations of the candidate MIFGMOSFET obtained by using the 90 nm SPICE BSIM4 based Monte-Carlo simulation with 99% confidence.
By :Banchuin Rawid (Department of Computer Engineering, Siam University, Bangkok, Thailand)、Chaisricharoen Roungsan (School of Information Technology, Mae Fah Luang University, Chiangrai, Thailand)
Link : jglobal.jst.go.jp/public
2. Forecasting Foreign Exchange Rate Using Time Series Analysis with Data Mining Techniques
The purpose of this research was to develop the model of forecasting foreign exchange rate in Thailand in order to make the planning decisions for business sectors and country economic development. This research provided three techniques of Time Series Data Mining Analysis, namely Linear Regression, Multi-Layer Perceptron and Support Vector Machine for Regression. The data used for the study collected from the foreign exchange rate from 2009 to 2017 AD. totally 108 months. The research findings showed that the suitable forecasting models for foreign exchange rate are US Dollar, Pound Stering, Euro, Yen and Yuan Renminbi. The forecasting model for foreign exchange rate which had the most accuracy were as followed: 1) The forecasting model using Support Vector Machine for Regression was the most suitable for US Dollar, Euro and Yen, which had the highest accuracy rate of MMRE (Mean Magnitude of Relative Error) with the percentage of 2.43, 1.39, and 2.57, respectively. 2) The forecasting model using Linear Regression was the most suitable for Pound Stering, which had the highest accuracy rate of MMRE with the percentage of 0.64. 3) The forecasting model using Multi-Layer Perceptron was the most suitable for Yuan Renminbi, which had the highest accuracy rate of MMRE with the percentage of 0.97.
KEYWORDS: Forecasting, Time Series Analysis, Data Mining Technique, Foreign Exchange Rate
Title: Forecasting Foreign Exchange Rate Using Time Series Analysis with Data Mining Techniques
By: Pitchayakorn Lake / Information Technology