KMITL

Permanent URI for this communityhttps://dspace.kmitl.ac.th/handle/123456789/1

Browse

Search Results

Now showing 1 - 8 of 8

Some of the metrics are blocked by your
consent settings
Time Series Forecasting Using Transfer Learning with an Attention-Revamped Transformer: A Case Study of Financial Instruments
(2025-12-01)
Feng, Ling
;
Sinchai, Ananta
Accurate financial time series forecasting is essential for informed investment and risk management decisions. Traditional methods, including statistical techniques such as SMA, ARIMA, VAR, and LASSO, and deep learning models like LSTM and Transformer, often fall short of capturing the complex seasonal and cyclical dynamics inherent in financial data. To overcome the noted shortcomings, this study introduces a transfer learning approach utilizing an enhanced Transformer model with correlation-based attention mechanisms. This proposed model significantly improves its capacity to capture long-term dependencies and perform robust cross-market predictions. Initially trained on the Dow Jones Index, it demonstrates superior transferability to diverse asset classes, including stock indices, commodities, and cryptocurrencies. Experimental evaluations across multiple metrics, including MSE, MAE, MHD, and R2, reveal that the proposed model consistently outperforms benchmarks. Notably, it achieves outstanding predictive accuracy in BTC (MSE: 0.0352) and SET (MSE: 0.1701), establishing a strong foundation for advanced transfer learning applications in financial forecasting across varied markets.
Some of the metrics are blocked by your
consent settings
Growth and yield performance of Bambara groundnut advanced line and introduced accessions grown in the rainy season in the South of Thailand
(2023-05-01)
Wongwichaiwat, S.
;
Chotechung, S.
;
Anothai, J.
;
Phakamas, N.
Result showed that the days to 50% flowering among four bambara groundnut genotypes ranged from 31.75 days to 34.25 days and days to harvest ranged from 93 days to 125 days. TVsu 89 had the lowest days to 50% flowering of 31.75 days and days to harvest of 93 days. TVsu 89 had lower agronomic traits than check variety (Songkhla 1), but it was not different from check variety for fresh pod yield and dry pod yield. Correlation analysis indicated that harvest index was positively and significantly associated with fresh pod yield and dry pod yield.
Some of the metrics are blocked by your
consent settings
Mobility retardation of Cd, Pb and Mn in acid soil using phosphate fertilizers
(2019-05-01)
Chaiyaraksa, Chompoonut
;
Rodsa, Natthanan
Contamination of heavy metals in soil is a major problem that causes damage to the environments. The aim of this research was to observe the efficiency of phosphate fertilizer, including phosphate rock, di-ammonium phosphate, and monopotassium phosphate in stabilizing lead, cadmium, and manganese in contaminated soil. The sampling soil was an acid sandy clay loam soil from Rayong Province which is one of the most industrialized provinces in Thailand. After applying fertilizers to the soil, the determination was on soil pH, the total concentration of heavy metals, heavy metal forms in soil and potential of heavy metals to enter the biological system. The results showed that phosphate rock, diammonium phosphate, and monopotassium phosphate increased the soil pH from 3.60 to 6.5, 7.0 and 5.2, respectively. Phosphate fertilizers could change an unstable form to a more stable form of heavy metals. Phosphate rock (7.5 g/kg<inf>soil</inf>) has the highest potential for reducing the mobility of all three metals (about 80% for manganese, 60% for cadmium, and 50% for lead), followed by monopotassium phosphate. The results obtained from the extraction with DTPA and CaCl<inf>2</inf> were closely related to the results obtained from the sequential extraction method. Phosphate rock was the best to reduce potentially toxic metals phytoavailability. Soil improvement with phosphate fertilizer was considered a good alternative for stabilizing soils contaminated with cadmium, lead, and manganese.
Some of the metrics are blocked by your
consent settings
A study of correlation for reverse logistics barriers to solutions using structural equation modelling
(2019-01-01)
Sirisawat, Pornwasin
;
Kiatcharoenpol, Tossapol
Reverse logistics (RL) plays an important role in many industries especially in the electronics industry and helps sustainability businesses while also reducing environmental impacts. The electronics industry is one of the most rapidly growing industries in Thailand. Many companies adopted RL for their business process but still enforce some barriers that make it unsuccessful. Therefore, this study aims to investigate the relationship of RL barriers and solutions then examines the relationship between these factors. A conceptual model was developed by proposing four hypotheses in which structural equation modelling (SEM) was used to examine the proposed model. The main results of this study show that there are statistics significant of the proposed hypothesis. In conclusion, the proposed model could fit with the empirical data. The proposed results could reconfirm that there are statistics significant of the relationship between RL barriers and solutions of RL practices implementation in Thailand electronics industry.
Some of the metrics are blocked by your
consent settings
Enhanced hydrocarbon production and improved biodiesel qualities of Botryococcus braunii KMITL 5 by vitamins thiamine, biotin and cobalamin supplementation
(2018-01-01)
Ruangsomboon, Suneerat
;
Sornchai, Piyanuch
;
Prachom, Noratat
Botryococcus braunii, a green microalga which is a potentially good biodiesel feedstock, was isolated from a freshwater fish pond in Bangkok, and the effects of B-vitamins—thiamine, biotin and cobalamin as well as their mixes—on its biomass, hydrocarbon production, carbohydrate content and biodiesel properties were investigated, using a control group that was not supplemented with any vitamins. The addition of thiamine increased the alga's biomass and hydrocarbon content, whereas addition of biotin, cobalamin and mixes of all three vitamins did not increase its hydrocarbon content. Among the vitamins tested, thiamine 150 μg L− 1 was the best for producing high biomass (2.08 ± 0.07 g L− 1), hydrocarbon content (37.29 ± 0.41%), hydrocarbon yield (0.77 ± 0.03 g L− 1) and hydrocarbon productivity (90.32 ± 3.05 mg L− 1 d− 1). Maximum carbohydrate content (386 ± 39 mg g− 1) was achieved by addition of 1 μg L− 1 cobalamin. Compared to the control group, the alga cultured in media supplemented with 100 μg L− 1 thiamine showed better biodiesel properties with lower iodine value (76.42 g I<inf>2</inf> 100 g− 1), cold filter plugging point (9.90 °C) and higher cetane number (54.30). Therefore, thiamine was shown to be a promising culture medium supplement for this strain for using it as a feedstock for biodiesel production. An investigation of vitamin B-related gene in B. braunii KMITL 5 found related genes for thiamine (thi), biotin (bioB and bioF) and cobalamin (metE and metH) vitamins, which are marks of vitamin autotrophy in algae.
Some of the metrics are blocked by your
consent settings
Predicting SET50 stock prices using CARIMA (Cross Correlation ARIMA)
(2016-02-08)
Wichaidit, Sornpon
;
Kittitornkun, Surin
Investing in stocks is one of the most popular approaches for money investment. This paper aims to predict short-term stock prices of SET50 of Stock Exchange of Thailand (SET). The proposed method is called CARIMA (Cross Correlation Autoregressive Integrated Moving Average. The basic idea of CARIMA is to find the most highly correlated s tock t o predict the target one in addition to ARIMA predicted price. The results of CARIMA model yield better price trends (measured by 10-day correlation coefficient) while % MAEs (Mean Absolute Errors) are quite similar with those of ARIMA.
Some of the metrics are blocked by your
consent settings
A new feature selection based on class dependency and feature dissimilarity
(2015-11-20)
Claypo, Niphat
;
Jaiyen, Saichon
Feature selection method is an important task for data preprocessing in data mining. Before a classifier learns the training data, there are a lot of features in each data set that makes the learning process slower. It is not appropriated for big data analytics. This paper proposes feature selection method based on the class dependency and feature dissimilarity (CDFD) using mutual information and Euclidean distance. The mutual information is applied to determine the dependency between the feature and the class if the dataset contains discrete data. If the dataset contains continuous data, the correlation between the feature and the class is used instead. The Euclidean distance is used for reducing the duplicated features based on dissimilarity between features. The experiments are conducted on five datasets. From the experimental results, the propose feature selection method can reduce the number of features in the data set and reduce the classification error of classifiers. Furthermore, it can be applied to discrete and continuous data and it can help classifiers improving their classification accuracies and reducing the computational times for learning.
Some of the metrics are blocked by your
consent settings
An empirical decomposition of deep groundwater time series and possible link to climate variability
(2014-01-01)
Seeboonruang, U.
Deep groundwater data reflects hydrological processes, climate change and variability, as well as any anthropogenic influence. Decomposition of deep groundwater signal examines the history of the groundwater region. Detrending is a vital step in decomposition of groundwater time series because it is expected to remove anthropogenic effects and long-term cyclic patterns. Eight detrending methods were applied to long-term groundwater records monitored in the Lower Chao Phraya basin in Thailand. Detrended residuals and subsequently periodograms of the residuals were computed by applying the Fourier series analysis. The result from this study indicates that the 5th order polynomial interpolation provides the trendlines that significantly relate to the groundwater withdrawal background. The detrended residual function is imbedded with two major cyclic patterns, which can be the result from global climate variability, e.g. Indian Ocean Dipole and the El Niño Southern Oscillation. The magnitude of deep groundwater dynamics as the result from the anthropogenic effect, is much greater than that of the climate variability in this region. In addition, this study demonstrates that caution must be exercised when fitting groundwater time series with different detrending techniques can yield mistaken cyclic patterns and may infer to different climate variability phenomenon. © 2014 Global NEST Printed in Greece. All rights reserved.

KMITL

Browse

Filters

Settings

Sort By

Results per page

Search Results