Chaiwuttisak, Pornpimol
Loading...
11 results
Now showing 1 - 10 of 11
- Some of the metrics are blocked by yourconsent settings
Item type:Publication, Machine Learning Based Decision Support System for High-School Study(2021-01-01)The objectives of this study are to investigate the correlations between personal factors, learning factors family, and economic factors affecting high-school study program selection and also to create and compare models of high-school study program selection with data mining techniques and to develop a decision support system for high-school study program selection with a data mining technique. Data were analyzed by five data mining techniques, and models of high-school study program selection were constructed. These models were then used to construct a decision support system from data mining software called RapidMiner Studio 9. The research findings were as follows personal factors, learning factors family, and economic factors affecting high-school study program selection, and from the result of high-school study program selection, the Decision Tree method, C4.5 algorithm provided the highest accuracy. Therefore, the researcher selected the forecasting model with the Decision Tree method, C4.5 algorithm together with the selection of features with the backward elimination method to create a decision support system. - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Applying Convolutional Neural Network with Candlestick Images to Predict Corn Price Movement(2025-01-01)The objective of this research is to study the effect of hyperparameters on corn price movement prediction models, namely batch size and learning rate, and create a model to predict the corn price movement in the Chicago Board of Trade (CBOT) based on candlestick images at 5-day and 20-day timeframes. The data are split into three sets, namely, training set, validation set, and test set, with a ratio of 70:10:20. The models presented in this research are CNN, VGG-16, and Efficientnet-B0, which must be fine-tuned. The study’s findings on hyperparameter values within a 5-day timeframe revealed that the optimal batch size and learning rates for all three models were a batch size of 16 with a learning rate of 0.001 and a timeframe of 20 days with a dataset size of 16. However, the suitable learning rate for the CNN model was 0.001, while for the VGG-16 and EfficientNet-B0 models, it was 0.0001. Subsequently, the hyperparameter values were fine-tuned for each model and tested the model with the test set. The study findings revealed that at the 5-day timeframe, the customized CNN model outperformed other models in predicting corn price movement, with an accuracy of 55.39%, while at a 20- day timeframe, the model with the highest accuracy was EfficientNet-B0, with an accuracy of 55.03%. - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Predicting the Workload in Debt Collections to Improve the Efficiency of Manpower Planning(2023-01-01)This research aimed to predict volume workloads in debt collections and compare the accuracy of forecasting methods. Three forecasting methods were considered in the study: the SARIMA model, the random forest model, and long short-term memory. The data used in this study were a time series of a daily debt collection volume workload due 1, 5, 10, 15, 20, and 25 in the used car loan corporation. They were divided into 2 datasets. The first dataset containing the past data from June 2020 to October 2020 was used for selecting the most suitable model, and the second dataset containing the past data from November 2020 to December 2020 was used for comparing the prediction accuracy of each forecasting model in terms of mean absolute percentage error (MAPE). Consequently, the lowest MAPE achieved by LSTM for different due dates was 6.79%, 6.24%, 7.08%, 10.88%, 12.45%, 10.17%, respectively, while the MAPE achieved by random forest was 11.72%, 11.27%, 8.94%, 11.92%, 13.91%, 16.67%, respectively, and the MAPE achieved by SARIMA was 23.35%, 32.87%, 32.58%, 30.07%, 33.93%, 27.31%, respectively. It indicates that the LSTM model was the most accurate to forecast the daily debt collection volume workload of the field debt collector in advance of the future. The predicted workload can be used as a piece of supporting information for adequate workforce planning of the corporation’s used car loan. - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Applying an Improved Ant Colony Optimization to solve the Homogeneous Fixed Fleet Close Open Mixed Vehicle Routing Problem(2021-04-01) - Some of the metrics are blocked by yourconsent settings
Item type:Publication, A Comparative Study of Machine Learning and Deep Learning Models for Technology Stock Price Prediction Using News Sentiment and Economic Indicators(2026-01-01)This study developed predictive models for the closing prices of five leading technology stocks: GOOGL, MSFT, AAPL, NVDA, and META by employing five advanced machine learning and deep learning techniques: Light Gradient Boosting Machine (LightGBM), Extreme Gradient Boosting (XGBoost), Recurrent Neural Network (RNN), Gated Recurrent Unit (GRU), and Long Short-Term Memory (LSTM). The modeling framework integrated sentiment scores derived from financial news articles specific to each stock using the VADER Sentiment Analysis tool, in conjunction with a range of macroeconomic indicators. Model performance was evaluated separately for each stock using Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE) as primary metrics. To determine whether statistically significant differences existed among the predictive performance of the models across all stocks, the Friedman test was employed, followed by the Wilcoxon signed-rank test for post-hoc pairwise comparisons. The empirical results indicated that XGBoost achieved superior predictive accuracy for MSFT and AAPL, GRU outperformed other models for NVDA and META, while RNN yielded the most accurate forecasts for GOOGL. - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Latent topic analysis of the post property for sales to predict a selling price of second-hand condominiums(2021-10-22)This research objective is to study the latent topics analysis in selling post of real estate of second-hand condominium by using Latent Dirichlet Allocation (LDA) and build a price prediction model of second-hand condominium using multiple linear regression and artificial neural networks by measuring and comparing the performance of the second hand condominium price prediction model with root mean square error (RMSE). This experiment included four variables are room size, number of bathroom, number of bedroom and latent topics from LDA. The result of LDA indicated that selling post of real estate can be separated into 4 topics, in which finding the factors that affect the price use the regression analysis method to get five variables are room size, number of bathroom, floors, topic 2 and topic 4. The RMSE based on the multiple linear regression analysis was 1.349, while the RMSE based on artificial neural network was 1.156. Thus, it can be concluded that the predictive model using the artificial neural networks is superior to multiple linear regression. - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Artificial Intelligence Application in Automated Odometer Mileage Recognition of Freight Vehicles(2021-01-01)Transportation cost management is necessary for entrepreneurs in industry and business. One way to do this is to report the daily mileage numbers read by the employees of a company, but still encountering errors in human mileage reading, resulting in incorrect information received and difficulties to plan effective revenue management. As well as, increasing workload and creating complications for employees in checking mileage information. Therefore, the objective of this research is to create a machine learning model for detecting and reading the mileage numbers 2 types of freight vehicles: Analog and Digital. It can be divided into 2 parts: 1) detect mileage is used to identify the position of the mileage in the image and cut only the mileage by removing the unrelated background from the image 2) detect numbers and reads miles. Both use object detection with the Faster-RCNN. The results show that to detect the position of the miles and cut only the number of miles to read the numbers correctly, 187 images from 220 test images, which is correct for the model, representing 85%. The results of the study achieve satisfactory performance that meet the requirements needed for real-life applications in the transportation and logistics industry. - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Multi-Criteria Decision-Making for Investment Portfolio Selection in Thailand's Stock Market(2023-07-21)This research aimed to determine the relative importance of four criteria for decision-making by using each one with a popular decision-making method and evaluating the outcomes and using these criteria to provide ranked alternatives (according to each criterion's relative importance) for making an investment decision. The four criteria were obtained from a comprehensive literature review related to securities investment. The investment data analyzed were past investment data on trading securities under the Energy and Utilities category of the SET50 index in the Stock Exchange of Thailand. The analysis was done through an Analytic Hierarchy Process (AHP) and a Technique for Order of Preference by Similarity to the Ideal Solution (TOPSIS). Opinions of three experts with experience in giving securities investment advice were collected and arranged into pair-wise comparison matrices that were used in AHP. AHP and TOPSIS calculations were done in Microsoft Excel. The results of the study show that the most important criterion was financial fundamentals with a weight of 44.59%; the second rank criterion was technical factors with a weight of 20.15%; the third-rank criterion was risk factors with a weight of 19.64%; and the last rank criterion was fundamentals of structure and sustainable development with a weight of 15.62%. In addition, the outcome of security ranking by TOPSIS and the past security ranking data were significantly similar as analyzed by a hypothesis statistical test with two dependent samples. - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Text Mining Analysis of Comments in Thai Language for Depression from Online Social Networks(2021-01-01)The objectives of this research were to analyze the relationship of the phrases or words commonly found in the comments from depression hashtag on Twitter using the association rules. The data used in this study were collected from comments in Thai language via depression hashtag on Twitter during 1 January 2019 to 31 January 2019, in total of 1,500 comments. According to the comments in Thai language on social media collected by using Rapidminer Studio 9 software to get the word about depression and used to analyze relationships of words from a text comment to get the format data Association. The frequency of words and phrases in a form of presentation is used to describe the various opinions about the depression that has a presentation on social media. According to the model performance in each of the above methods, it was found that Euclidean Distance provided the best result due to the smallest average distance at all points in each cluster which was equal to 152.504. The association analysis, a total of 30 association rules were obtained, the support of 0.5% and the 80% confidence. - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Analysis of Driver's Attention through the Internet of Things (IOTs) for Preventing Road Accident of Natural Gas Vehicles(2021-04-01)