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
Work-related generative artificial intelligence adoption and employee-perceived organisational performance in Thailand's manufacturing industry
(2026-10-01)
Yudhistyra, Wecka Imam
;
Srinuan, Chalita
Generative Artificial Intelligence (GenAI) is increasingly integrated into organisational work processes, yet its implications for manufacturing firms cannot be inferred from adoption intention or usage frequency alone. In contrast to deterministic workplace technologies, GenAI produces probabilistic and context-responsive outputs whose organisational value depends on employees' ability to evaluate, interpret, and apply generated content effectively. This research examines how work-related GenAI use is associated with employee perceptions of GenAI-enabled organisational performance in Thailand's manufacturing sector. Drawing on the Unified Theory of Acceptance and Use of Technology (UTAUT) framework, the Human-Centred Design paradigm, and Knowledge Management theory, the research develops a framework in which perceived usability, knowledge acquisition, knowledge application, and performance expectancy influence behavioural intention, which subsequently affects use behaviour and employee-perceived GenAI-enabled organisational performance. Survey data from 381 manufacturing employees with prior GenAI experience were analysed using Partial Least Squares Structural Equation Modelling. Bayesian Network (BN) algorithms were subsequently employed to refine the theoretical structure by identifying conditional dependencies beyond the hypothesised model, while Artificial Neural Networks and Importance-Performance Map Analysis provided complementary predictive and managerial insights. The results support all hypothesised relationships. The BN-refined model further identifies direct links between knowledge application, performance expectancy, and organisational performance, suggesting that, in addition to adoption and use, GenAI-derived value also depends on employees' capacity to translate AI-generated knowledge into workplace outcomes. The findings advance technology adoption research by developing and validating a GenAI-specific extension of the UTAUT that integrates human-centred design and knowledge management perspectives. The research further demonstrates that organisational value is realised through a sequential pathway linking adoption antecedents, behavioural intention, and use behaviour, while providing policy-relevant evidence on workforce-driven AI adoption in an emerging economy undergoing digital transformation.
Some of the metrics are blocked by your
consent settings
Localized estimation of event-related neural source activity from simultaneous MEG-EEG with a recurrent neural network
(2024-12-01)
O'Reilly, Jamie A.
;
Zhu, Judy D.
;
Sowman, Paul F.
Estimating intracranial current sources underlying the electromagnetic signals observed from extracranial sensors is a perennial challenge in non-invasive neuroimaging. Established solutions to this inverse problem treat time samples independently without considering the temporal dynamics of event-related brain processes. This paper describes current source estimation from simultaneously recorded magneto- and electro-encephalography (MEEG) using a recurrent neural network (RNN) that learns sequential relationships from neural data. The RNN was trained in two phases: (1) pre-training and (2) transfer learning with L1 regularization applied to the source estimation layer. Performance of using scaled labels derived from MEEG, magnetoencephalography (MEG), or electroencephalography (EEG) were compared, as were results from volumetric source space with free dipole orientation and surface source space with fixed dipole orientation. Exact low-resolution electromagnetic tomography (eLORETA) and mixed-norm L1/L2 (MxNE) source estimation methods were also applied to these data for comparison with the RNN method. The RNN approach outperformed other methods in terms of output signal-to-noise ratio, correlation and mean-squared error metrics evaluated against reference event-related field (ERF) and event-related potential (ERP) waveforms. Using MEEG labels with fixed-orientation surface sources produced the most consistent estimates. To estimate sources of ERF and ERP waveforms, the RNN generates temporal dynamics within its internal computational units, driven by sequential structure in neural data used as training labels. It thus provides a data-driven model of computational transformations from psychophysiological events into corresponding event-related neural signals, which is unique among MEEG source reconstruction solutions.
Some of the metrics are blocked by your
consent settings
Transient staiblity analysis by adaptive neuro fuzzy inference system and sobol sequence
(2018-07-02)
Phootrakornchai, Witsawa
;
Jiriwibhakorn, Somchat
It is known that the time domain is an accurate method that is used for the assessment of transient stability and critical clearing time for any power systems. However, the time domain method normally takes a long time for the calculation due to many differential and non-linear equations, thus it may not be appropriate to apply with the real-time analysis, especially the large power system. We try to find any approaches to minimize the computation time and maximize the accuracy of results as much as possible. This paper therefore proposes a method using adaptive neuro fuzzy inference system and sobol sequence for solving the critical clearing time. The approach proves that it can give us the satisfactory prediction of critical clearing time even for a large power system. The result obtained by adaptive neuro fuzzy inference system are also compared to the result obtained by means of artificial neural network being generally used for the power system analysis.
Some of the metrics are blocked by your
consent settings
Heart disease classification using neural network and feature selection
(2011-11-04)
Khemphila, Anchana
;
Boonjing, Veera
In this study, we introduces a classification approach using Multi-Layer Perceptron (MLP)with Back-Propagation learning algorithm and a feature selection algorithm along with biomedical test values to diagnose heart disease. Clinical diagnosis is done mostly by doctor's expertise and experience. But still cases are reported of wrong diagnosis and treatment. Patients are asked to take number of tests for diagnosis. In many cases, not all the tests contribute towards effective diagnosis of a disease. Our work is to classify the presence of heart disease with reduced number of attributes. Original, 13 attributes are involved in classify the heart disease. We use Information Gain to determine the attributes which reduces the number of attributes which is need to be taken from patients. The Artificial neural networks is used to classify the diagnosis of patients. Thirteen attributes are reduced to 8 attributes. The accuracy differs between 13 features and 8 features in training data set is 1.1% and in the validation data set is 0.82%. © 2011 IEEE.
Some of the metrics are blocked by your
consent settings
Comparing performances of logistic regression, decision trees, and neural networks for classifying heart disease patients
(2010-12-01)
Khemphila, Anchana
;
Boonjing, Veera
In this study, performances of classification techniques were compared in order to predict the presence of the patients getting a heart disease.A retrospective analysis was performed in 303 subjects.We compared the performance of logistic regression(LR),decision trees(DTs) , and Artificial neural networks (ANNs).The variables were medical profiles are age,Sex,Chest Pain Type,Blood Pressure,Cholesterol,Fasting Blood Sugar, Resting ECG,Maximum Heart Rate,Induced Angina,Ole Peak,Slope,Number Colored Vessels,Thal and Concept Class.We have created the model using logistic regression classifiers , artificial neural networks and decision trees that they are often used for classification problems.Performances of classification techniques were compared using lift chart and error rates.In the result, artificial neural networks have the greatest area between the model curve and the baseline curve.The error rates are 0.22,0.198,0.21,respectively for logistic regression , artificial neural networks and decision trees.The neural networks exhibited sensitivity of 81.1% , specificity of 78.7% and accuracy of 80.2%,while the decision tree provided the prediction performance with a sensitivity, specificity and accuracy of 81.7%,76.0% and 79.3%.And the logistic regression provided the prediction performance with a sensitivity,specificity and accuracy of 81.2%,73.1% and 77.7% Artificial neural networks have the least of error rate and has the highest accuracy ,therefore Artificial neural networks is the best technique to classify in this data set. ©2010 IEEE.
Some of the metrics are blocked by your
consent settings
Artificial neural networks for predicting the maximum surface settlement caused by EPB shield tunneling
(2006-03-01)
Suwansawat, Suchatvee
;
Einstein, Herbert H.
Numerous empirical and analytical relations exist between shield tunnel characteristics and surface and subsurface deformation. Also, 2-D and 3-D numerical analyses have been applied to such tunneling problems. Similar but substantially fewer approaches have been developed for earth pressure balance (EPB) tunneling. In the Bangkok MRTA project, data on ground deformation and shield operation were collected. The tunnel sizes are practically identical and the subsurface conditions over long distances are comparable, which allow one to establish relationships between ground characteristics and EPB - Operation on the one hand, and surface deformations on the other hand. After using the information to identify which ground- and EPB-characteristic have the greatest influence on ground movements, an approach based on artificial neural networks (ANN) was used to develop predictive relations. Since the method has the ability to map input to output patterns, ANN enable one to map all influencing parameters to surface settlements. Combining the extensive computerized database and the knowledge of what influences the surface settlements, ANN can become a useful predictive method. This paper attempts to evaluate the potential as well as the limitations of ANN for predicting surface settlements caused by EPB shield tunneling and to develop optimal neural network models for this objective. © 2005 Elsevier Ltd. All rights reserved.
Some of the metrics are blocked by your
consent settings
Determining the orders of feature and hidden unit prunings of artificial neural networks
(2005-12-01)
Jearanaitanakij, Kietikul
;
Pinngern, Ouen
There is a great deal of research undertaken for pruning away features and hidden units in order to reduce the size of Artificial Neural Networks (ANNs). However, none of these methods mentions about the relationship between the pruned unit and the number of epochs needed for retraining when the unit is pruned away from the network. In this paper, we present two heuristics for determining the pruning orders, which lead to the near smallest number of retraining epochs. The heuristics are based on the employment of the modified information gain calculated from all features in training data. Then, we test our proposed heuristics on an exclusive-or data set. The experimental results show the success of using information gain as a criterion for determining the pruning orders. © 2005 IEEE.
Some of the metrics are blocked by your
consent settings
Automatic indexing system for atmospheric laser radar data
(2002-01-01)
Lerkvaranyu, S.
;
Miyanaga, Y.
;
Dejhan, K.
;
Cheevasuvit, F.
;
Mizutani, K.
The purpose of this paper is to design a new method for an automatic indexing system with unsupervised conditions. In this paper, the method of a self-organizing clustering network is adopted. It is used to classify and index a large amount of real atmospheric laser radar data. Initially, the parameters of each cluster will start with random initial values and are adapted with the algorithm. In this paper, six groups are clustered from the given data. It is also shown that some of these indicate quite important atmospheric condition characteristics.

KMITL

Browse

Filters

Settings

Sort By

Results per page

Search Results