Pub Date : 2022-01-26DOI: 10.1109/kst53302.2022.9729074
{"title":"[4 Messages]","authors":"","doi":"10.1109/kst53302.2022.9729074","DOIUrl":"https://doi.org/10.1109/kst53302.2022.9729074","url":null,"abstract":"","PeriodicalId":433638,"journal":{"name":"2022 14th International Conference on Knowledge and Smart Technology (KST)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-01-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125188827","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Deep learning is a technique for image processing and data analysis with promising results and large potential. We investigated the performance of Deep Convolutional Neural Network (DCNN) for recognizing our spatiotemporal data in surveillance camera images. We studied how the magnitude of image dataset affected DCNN base models. We extracted spatialtemporal data into seven different interval datasets of Okra vegetation and applied them to two well-known convolutional networks; AlexNet and GoogLeNet. We experimented with spatiotemporal datasets on the convolutional networks and compared them in different epochs. The 1-Minute, 15-Minute, and 30-Minute periodic spatiotemporal datasets can achieve an excellent deep learning model with accuracy higher than 99% for both AlexNet and GoogLeNet.
{"title":"A Deep Learning-Based Spatial and Temporal Data: Plant-Growing Case Study","authors":"Barakatullah Azizi, Narongrit Waraporn, Murray Leigh Ayres","doi":"10.1109/KST53302.2022.9729064","DOIUrl":"https://doi.org/10.1109/KST53302.2022.9729064","url":null,"abstract":"Deep learning is a technique for image processing and data analysis with promising results and large potential. We investigated the performance of Deep Convolutional Neural Network (DCNN) for recognizing our spatiotemporal data in surveillance camera images. We studied how the magnitude of image dataset affected DCNN base models. We extracted spatialtemporal data into seven different interval datasets of Okra vegetation and applied them to two well-known convolutional networks; AlexNet and GoogLeNet. We experimented with spatiotemporal datasets on the convolutional networks and compared them in different epochs. The 1-Minute, 15-Minute, and 30-Minute periodic spatiotemporal datasets can achieve an excellent deep learning model with accuracy higher than 99% for both AlexNet and GoogLeNet.","PeriodicalId":433638,"journal":{"name":"2022 14th International Conference on Knowledge and Smart Technology (KST)","volume":"353 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-01-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122785813","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-01-26DOI: 10.1109/KST53302.2022.9729084
A. Rahman, Muntequa Imtiaz Siraji, Lamim Ibtisam Khalid, Fahim Faisal, M. M. Nishat, Mohammad Tausiful Islam, Nchouwat Ndumgouo Ibrahim moubarak
The mental state of a person is a combination of very complex neural activities which determine the current state of mind. It depends on a lot of external factors as well as internal factors of the brain itself. It is possible to determine an individual's mental state by analyzing their EEG patterns. Using a dataset acquired from Kaggle, ten machine learning techniques were investigated and models were built. The RandomSearchCV method was used to perform hyperparameter tuning and a comparative study has been portrayed for both tuning and without tuning of hyperparameter. After evaluating the performance parameters, Support Vector Machine (SVM) displayed the best accuracy (95.36%). However, Gradient Boosting (GrB) depicted promising accuracy of 95.24% whereas K-Nearest Neighbors (KNN) and XGBoost (XGB) both depicted 93.10% accuracy. As a result, with effective integration of the ML-based detection method, it is likely to regulate a person's state of mind, which will enable to develop a better understanding of human psychology and forecast their actions.
{"title":"Detection of Mental State from EEG Signal Data: An Investigation with Machine Learning Classifiers","authors":"A. Rahman, Muntequa Imtiaz Siraji, Lamim Ibtisam Khalid, Fahim Faisal, M. M. Nishat, Mohammad Tausiful Islam, Nchouwat Ndumgouo Ibrahim moubarak","doi":"10.1109/KST53302.2022.9729084","DOIUrl":"https://doi.org/10.1109/KST53302.2022.9729084","url":null,"abstract":"The mental state of a person is a combination of very complex neural activities which determine the current state of mind. It depends on a lot of external factors as well as internal factors of the brain itself. It is possible to determine an individual's mental state by analyzing their EEG patterns. Using a dataset acquired from Kaggle, ten machine learning techniques were investigated and models were built. The RandomSearchCV method was used to perform hyperparameter tuning and a comparative study has been portrayed for both tuning and without tuning of hyperparameter. After evaluating the performance parameters, Support Vector Machine (SVM) displayed the best accuracy (95.36%). However, Gradient Boosting (GrB) depicted promising accuracy of 95.24% whereas K-Nearest Neighbors (KNN) and XGBoost (XGB) both depicted 93.10% accuracy. As a result, with effective integration of the ML-based detection method, it is likely to regulate a person's state of mind, which will enable to develop a better understanding of human psychology and forecast their actions.","PeriodicalId":433638,"journal":{"name":"2022 14th International Conference on Knowledge and Smart Technology (KST)","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-01-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121641172","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-01-26DOI: 10.1109/KST53302.2022.9729082
Alexandros I. Metsai, Konstantinos Karamitsios, Konstantinos Kotrotsios, P. Chatzimisios, G. Stalidis, Kostas Goulianas
Recommender Systems are a highly active area in research and development that has taken advantage of the recent progress in artificial intelligence and deep learning algorithms. Collaborative Filtering approaches have utilized neural networks for modelling complex nonlinear relationships regarding the interactions of users and items, with numerous commercial platforms utilizing such systems for providing personalized recommendations to their users. In this work, we present the evolution of the field and the most influential approaches, from simpler neural network models expanding matrix factorization techniques, to increasingly more complex architectures. We report notable applications and highlight key differences between research and production settings. At the same time, we note that the evaluation approaches followed by the literature vary, and we underline the significance of testing models during production with methods such as A/B tests and the measurement of key performance indicators, aside from offline testing.
{"title":"Evolution of Neural Collaborative Filtering for Recommender Systems","authors":"Alexandros I. Metsai, Konstantinos Karamitsios, Konstantinos Kotrotsios, P. Chatzimisios, G. Stalidis, Kostas Goulianas","doi":"10.1109/KST53302.2022.9729082","DOIUrl":"https://doi.org/10.1109/KST53302.2022.9729082","url":null,"abstract":"Recommender Systems are a highly active area in research and development that has taken advantage of the recent progress in artificial intelligence and deep learning algorithms. Collaborative Filtering approaches have utilized neural networks for modelling complex nonlinear relationships regarding the interactions of users and items, with numerous commercial platforms utilizing such systems for providing personalized recommendations to their users. In this work, we present the evolution of the field and the most influential approaches, from simpler neural network models expanding matrix factorization techniques, to increasingly more complex architectures. We report notable applications and highlight key differences between research and production settings. At the same time, we note that the evaluation approaches followed by the literature vary, and we underline the significance of testing models during production with methods such as A/B tests and the measurement of key performance indicators, aside from offline testing.","PeriodicalId":433638,"journal":{"name":"2022 14th International Conference on Knowledge and Smart Technology (KST)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-01-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129128576","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-01-26DOI: 10.1109/KST53302.2022.9729083
Phenphitcha Patthanajitsilp, P. Chongstitvatana
This research aims to present the detection system of an obstacle for electric wheelchair using computer vision in order to facilitate for disabled persons and reduce the possibilities of accidents. In this system, the distance threshold is set to alert when a wheelchair is approaching an obstacle. The alert system consists of the smartphone's camera attached to the back of a wheelchair. The YOLOv3 model was used for object detection. The researcher has developed an algorithm to detect obstacles such as pillars, doors, or edge of the wall with edge detection method to enhance the detection efficiency of the system. Therefore, the usage of two algorithms enables the system to choose the obstacle detection between objects and edge detection. The research found that the system can choose the algorithm to detect obstacles with an accuracy of up to 80%. Moreover, the experiment revealed that the system can alert warnings before collisions with an accuracy of up to 90%. Further, this system can also calculate the approximate time prior to the collision.
{"title":"Obstacles Detection for Electric Wheelchair with Computer Vision","authors":"Phenphitcha Patthanajitsilp, P. Chongstitvatana","doi":"10.1109/KST53302.2022.9729083","DOIUrl":"https://doi.org/10.1109/KST53302.2022.9729083","url":null,"abstract":"This research aims to present the detection system of an obstacle for electric wheelchair using computer vision in order to facilitate for disabled persons and reduce the possibilities of accidents. In this system, the distance threshold is set to alert when a wheelchair is approaching an obstacle. The alert system consists of the smartphone's camera attached to the back of a wheelchair. The YOLOv3 model was used for object detection. The researcher has developed an algorithm to detect obstacles such as pillars, doors, or edge of the wall with edge detection method to enhance the detection efficiency of the system. Therefore, the usage of two algorithms enables the system to choose the obstacle detection between objects and edge detection. The research found that the system can choose the algorithm to detect obstacles with an accuracy of up to 80%. Moreover, the experiment revealed that the system can alert warnings before collisions with an accuracy of up to 90%. Further, this system can also calculate the approximate time prior to the collision.","PeriodicalId":433638,"journal":{"name":"2022 14th International Conference on Knowledge and Smart Technology (KST)","volume":"265 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-01-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133836613","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-01-26DOI: 10.1109/KST53302.2022.9729059
Dermot O'Brien, Vasileios Christaras, Ioannis Kounelis, I. N. Fovino, G. Fontaras
Vehicles are becoming increasingly connected and are already transmitting substantial amounts of data to the Original Equipment Manufacturers (OEMs) servers. Blockchain (BC) enables the transmission of data with addition security and removes single points of failure, while maintaining data prove-nance, identity ownership and the possibility to retain varying levels of privacy depending on requirements of the applied use-case. This research performs emulations of vehicles interacting with European Member State authorities and the European Commission (EC) BC nodes that are running Hyperlegder Fabric (HLF) and explores whether the technology is could be used for transport applications, building on indicative case studies such as CO2 emissions monitoring and vehicle identity. Due to the specialized nature of the Experimental Platform for Internet Contingency infrastructure, the network topology can be defined allowing for more realistic network conditions to be emulated. The results show that the deployed system is able to meet the requirements both in terms of Transactions Per Second and latency, but the hardware and system parameters need modification in order to scale up to the envisaged number of vehicles in the fleet.
{"title":"Blockchain for Transport (BC4 T), Performance Simulations of Blockchain Network for Emission Monitoring","authors":"Dermot O'Brien, Vasileios Christaras, Ioannis Kounelis, I. N. Fovino, G. Fontaras","doi":"10.1109/KST53302.2022.9729059","DOIUrl":"https://doi.org/10.1109/KST53302.2022.9729059","url":null,"abstract":"Vehicles are becoming increasingly connected and are already transmitting substantial amounts of data to the Original Equipment Manufacturers (OEMs) servers. Blockchain (BC) enables the transmission of data with addition security and removes single points of failure, while maintaining data prove-nance, identity ownership and the possibility to retain varying levels of privacy depending on requirements of the applied use-case. This research performs emulations of vehicles interacting with European Member State authorities and the European Commission (EC) BC nodes that are running Hyperlegder Fabric (HLF) and explores whether the technology is could be used for transport applications, building on indicative case studies such as CO2 emissions monitoring and vehicle identity. Due to the specialized nature of the Experimental Platform for Internet Contingency infrastructure, the network topology can be defined allowing for more realistic network conditions to be emulated. The results show that the deployed system is able to meet the requirements both in terms of Transactions Per Second and latency, but the hardware and system parameters need modification in order to scale up to the envisaged number of vehicles in the fleet.","PeriodicalId":433638,"journal":{"name":"2022 14th International Conference on Knowledge and Smart Technology (KST)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-01-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121068605","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-01-26DOI: 10.1109/KST53302.2022.9729077
Tipajin Thaipisutikul, T. Shih, Avirmed Enkhbat, Wisnu Aditya, H. Shih, P. Mongkolwat
With the restrictions in our daily life activities under the current situation of the covid-19 pandemic worldwide, billions of people rely on social media platforms to share and obtaining covid-19 related news information. This made social media platforms easily be used as a source of myths and disinformation, which can cause severe public risks. It is thus of vital importance to constraint the spread of misinformation to the public. Although many works have shown promising results on the misinformation detection problem, only a few studies focus on the infodemic detection during the covid-19 pandemic, especially in the low resource language like Thai. Therefore, in this paper, we conduct extensive experiments on the real-world social network datasets to detect misinformation about covid-19 targeting both English and Thai languages. In particular, we perform an exploratory data analysis to get the statistic and characteristics of real and fake content. Also, we evaluate a series of three feature extraction, seven traditional machine learning, and eleven deep learning methods in detecting the fabricated content on social media platforms. The experimental results demonstrate that the transformer-based model significantly outperforms other deep learning and traditional machine learning methods in all metrics, including accuracy and F-measure.
{"title":"Beyond fear go viral: A machine learning study on infodemic detection during covid-19 pandemic","authors":"Tipajin Thaipisutikul, T. Shih, Avirmed Enkhbat, Wisnu Aditya, H. Shih, P. Mongkolwat","doi":"10.1109/KST53302.2022.9729077","DOIUrl":"https://doi.org/10.1109/KST53302.2022.9729077","url":null,"abstract":"With the restrictions in our daily life activities under the current situation of the covid-19 pandemic worldwide, billions of people rely on social media platforms to share and obtaining covid-19 related news information. This made social media platforms easily be used as a source of myths and disinformation, which can cause severe public risks. It is thus of vital importance to constraint the spread of misinformation to the public. Although many works have shown promising results on the misinformation detection problem, only a few studies focus on the infodemic detection during the covid-19 pandemic, especially in the low resource language like Thai. Therefore, in this paper, we conduct extensive experiments on the real-world social network datasets to detect misinformation about covid-19 targeting both English and Thai languages. In particular, we perform an exploratory data analysis to get the statistic and characteristics of real and fake content. Also, we evaluate a series of three feature extraction, seven traditional machine learning, and eleven deep learning methods in detecting the fabricated content on social media platforms. The experimental results demonstrate that the transformer-based model significantly outperforms other deep learning and traditional machine learning methods in all metrics, including accuracy and F-measure.","PeriodicalId":433638,"journal":{"name":"2022 14th International Conference on Knowledge and Smart Technology (KST)","volume":"114 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-01-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117256849","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A trajectory of the tongue has several benefits in various domains such as articulatory and medical. It allows a user to analyze human speech or diagnose anomaly tongue movement of patients. This research focuses on estimating tongue motion. Most existing solutions apply traditional image processing techniques to a sequence of images to compute motion. Although they can precisely estimate a tongue motion, there are drawbacks to practicality and scalability. It is because of the high cost of medical imaging devices such as magnetic resonance imaging (MRI) and ultrasound scanners. There is also overhead in the preparation of marking on the face of the patient. On the other hand, the optical How algorithm can produce motion vectors on videos obtained from a commercial camera. This paper proposes a solution that can estimate tongue motion with more praetieality and less overhead. An average motion vector can be precisely computed within a region of interest of a tongne.
{"title":"Measurement of Tongue Motion using Optical Flows on Segmented Areas","authors":"Worapan Kusakunniran, Kittinun Aukkapinyo, Punyanuch Borwarnginn, Thanandon Imaromkul, Kittikhun Thongkanchorn, Disathon Wattanadhirach, Sophon Mongkolluksamee, Ratchainant Thammasudjarit, P. Ritthipravat, Pimchanok Tuakta, Paitoon Benjapornlert","doi":"10.1109/KST53302.2022.9729063","DOIUrl":"https://doi.org/10.1109/KST53302.2022.9729063","url":null,"abstract":"A trajectory of the tongue has several benefits in various domains such as articulatory and medical. It allows a user to analyze human speech or diagnose anomaly tongue movement of patients. This research focuses on estimating tongue motion. Most existing solutions apply traditional image processing techniques to a sequence of images to compute motion. Although they can precisely estimate a tongue motion, there are drawbacks to practicality and scalability. It is because of the high cost of medical imaging devices such as magnetic resonance imaging (MRI) and ultrasound scanners. There is also overhead in the preparation of marking on the face of the patient. On the other hand, the optical How algorithm can produce motion vectors on videos obtained from a commercial camera. This paper proposes a solution that can estimate tongue motion with more praetieality and less overhead. An average motion vector can be precisely computed within a region of interest of a tongne.","PeriodicalId":433638,"journal":{"name":"2022 14th International Conference on Knowledge and Smart Technology (KST)","volume":"45 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-01-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130589802","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-01-26DOI: 10.1109/KST53302.2022.9729078
Worawat Lawanont, Anantaya Timtong
Emerging technologies in the past decades have enabled many possibilities and higher education is no exception. Digital transformation in higher education has started many discussion from how to run a university to how to conduct a course. When looking at teach aspect specifically, it is mind blowing on the potential benefit the education system could have acquired if all data were put to the right application or system. With the support of various study on students traits and behaviors and how they affect their success, this study proposed an approach to harvest logged data from an online learning system of Suranaree University of Technology, then derived the learners' behaviors and used them as the dataset. The study developed total of five machine learning models to predict learners' score using the behavior data. The dataset used for the model training was related to the course progress. Thus, it was possible to predict the learners score as soon as the first week of the course. The results of this study shows promising accuracy, which can be used as a guideline approach to develop a decision support system to give immediate feedback to learners and resulting in transforming the way the learners learn.
{"title":"Smart Education Using Machine Learning for Outcome Prediction in Engineering Course","authors":"Worawat Lawanont, Anantaya Timtong","doi":"10.1109/KST53302.2022.9729078","DOIUrl":"https://doi.org/10.1109/KST53302.2022.9729078","url":null,"abstract":"Emerging technologies in the past decades have enabled many possibilities and higher education is no exception. Digital transformation in higher education has started many discussion from how to run a university to how to conduct a course. When looking at teach aspect specifically, it is mind blowing on the potential benefit the education system could have acquired if all data were put to the right application or system. With the support of various study on students traits and behaviors and how they affect their success, this study proposed an approach to harvest logged data from an online learning system of Suranaree University of Technology, then derived the learners' behaviors and used them as the dataset. The study developed total of five machine learning models to predict learners' score using the behavior data. The dataset used for the model training was related to the course progress. Thus, it was possible to predict the learners score as soon as the first week of the course. The results of this study shows promising accuracy, which can be used as a guideline approach to develop a decision support system to give immediate feedback to learners and resulting in transforming the way the learners learn.","PeriodicalId":433638,"journal":{"name":"2022 14th International Conference on Knowledge and Smart Technology (KST)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-01-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132758850","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-01-26DOI: 10.1109/KST53302.2022.9729076
Jakkaphan Whasphuttisit, Watchareewan Jitsakul, T. Kaewkiriya
There are researches that study about clustering techniques e.g., K-Means, K-Medoids, and X-Means. Their works mainly focus on applying one technique on multiple data sets to find the pros and cons of each algorithm. In this work, we focus on study and comparing these three clustering techniques instead. The experiment is done by applying each technique on Thai mutual funds fee data set which consists of 2,595 funds. From our experiment, we found that the optimal K value is 22. K-Means use the least processing time while K-Medoids use the most time. K-Means also has the least average distant between each centroid while K-Medoids has the most average distant. From Davies-Bouldin index, X-Means has the lowest value while K-Medoids has the highest value. The most density cluster of K-Means and X-Means is cluster 0 but it is cluster 1 for K-Medoids.
{"title":"Comparison of Clustering Techniques for Thai Mutual Funds Fee Dataset","authors":"Jakkaphan Whasphuttisit, Watchareewan Jitsakul, T. Kaewkiriya","doi":"10.1109/KST53302.2022.9729076","DOIUrl":"https://doi.org/10.1109/KST53302.2022.9729076","url":null,"abstract":"There are researches that study about clustering techniques e.g., K-Means, K-Medoids, and X-Means. Their works mainly focus on applying one technique on multiple data sets to find the pros and cons of each algorithm. In this work, we focus on study and comparing these three clustering techniques instead. The experiment is done by applying each technique on Thai mutual funds fee data set which consists of 2,595 funds. From our experiment, we found that the optimal K value is 22. K-Means use the least processing time while K-Medoids use the most time. K-Means also has the least average distant between each centroid while K-Medoids has the most average distant. From Davies-Bouldin index, X-Means has the lowest value while K-Medoids has the highest value. The most density cluster of K-Means and X-Means is cluster 0 but it is cluster 1 for K-Medoids.","PeriodicalId":433638,"journal":{"name":"2022 14th International Conference on Knowledge and Smart Technology (KST)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-01-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129008965","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：