Sunantha Sodsee’s research while affiliated with King Mongkut's University of Technology North Bangkok and other places

Artificial neural networks are just mathematical functions used for approximation purposes, and learning is a euphemism for determining the parameters of these functions. Experience from other areas of approximation theory shows that closed-form approximation functions determined globally such as neural networks tend to be quite rippled and are, thus, outperformed by approximation functions defined locally. With respect to artificial intelligence, neural networks lack explainability. Therefore, research in novel approaches for local approximators apt to replace neural networks and to outperform them with respect to accuracy, computational expense and explainability is suggested.KeywordsNeural networksuniversalglobal and local approximationartificial intelligenceexplainability

This book contains the proceedings of the 17th International Conference on Computing and Information Technology (IC2IT2021) that was held during May 13–14, 2021, in Bangkok, Thailand. The research contributions include machine learning, natural language processing, image processing, intelligent systems and algorithms, as well as network and cloud computing. These lead to the major research directions for emerging information technology and innovation, reflecting digital disruption in the world.

The number of air shipments with high added value is rapidly growing. For this kind of cargo, it is often necessary to monitor its specific characteristics also during transit. To enable this feature, a special kind of freight containers appropriate for air cargo ought to be employed, which can maintain prescribed micro-climatic conditions inside of them as well as communicate their parameters to their consignors. At the same time, a cooperative intelligent transport system should ensure valuable cargo’s protection against loss and damage throughout the entire transport chain. In fulfilment of these requirements an appropriate, safe and secure intelligent transport unit (iTU) is designed meeting the International Air Transport Association’s (IATA) standards for transporting pharmaceuticals as well as the principles of the Internet of Things and the arising Physical Internet. Deploying the iTU is expected to render handling and transportation of sensitive, high-value cargo by Cooperative Intelligent Transport Systems (C-ITS) more transparent, safe and efficient.

This book gathers the proceedings of the 16th International Conference on Computing and Information Technology (IC²IT 2020), held on May 14th–15th, 2020, at Dusit Thani Pattaya, Thailand. The topics covered include big data, artificial intelligence, machine learning, natural language processing, speech recognition, image and video processing, and deep learning. In turn, the topics represent major research and engineering directions for autonomous driving, language assistants, automatic translation, and answering systems. Lastly, they are responses to major economic changes around the world, which are increasingly shaped by the need for enhanced globalization and worldwide cooperation, and by emerging global problems.

As recommender systems have proven their effectiveness in providing personalised recommendations based on previous user preferences in e-commerce, this approach is to be transferred for use in medicine. In particular, the aim is to complement the diagnoses made by physicians in rural hospitals of developing countries, in remote areas or in situations of uncertainty by machine recommendations that draw on large bases of expert knowledge to reduce the risk to patients. To this end, a database of patients’ medical history and a cluster model is maintained centrally. The model is constructed incrementally by a combination of collaborative and knowledge-based filtering, employing a weighted similarity distance specifically derived for medical knowledge. In the course of this process, the model permanently widens its base of knowledge on a medical area given. To give a recommendation, the model’s cluster best matching the diagnostic pattern of a considered patient is sought. Fuzzy sets are employed to cope with possible confusion in decision making, which may occur when large data sets cause clusters to overlap. The degrees of membership to these fuzzy sets are expressed by the Mahalanobis distance, whose weights are derived from risk factors identified by experts. The therapy actually applied after the recommendation and its subsequently observed consequences are fed back for model updating. Readily available mobile digital accessories can be used for remote data entry and recommendation display as well as for communication with the central site. The approach is validated in the area of obstetrics and gynecology.

This book contains the research contributions presented at the 14th International Conference on Computing and Information Technology (IC2IT 2018) organised by King Mongkut’s University of Technology North Bangkok and its partners, and held in the northern Thai city of Chiang Mai in July 2018. Traditionally, IC2IT 2018 provides a forum for exchange on the state of the art and on expected future developments in its field. Correspondingly, this book contains chapters on topics in data mining, machine learning, natural language processing, image processing, networks and security, software engineering and information technology. With them, the editors want to foster inspiring discussions among colleagues, not only during the conference. It is also intended to contribute to a deeper understanding of the underlying problems as needed to solve them in complex environments and, beneficial for this purpose, to encourage interdisciplinary cooperation.

This research proposes the Top-k Items Recommendation System which uses clustering techniques based on memory-based collaborative filtering technique. Currently, data sparsity and quantity of system are problems in memory-based collaborative filtering technique. We offer recommend or show some items set for user’s preference. In this research, we propose methods for recommended items set to user preference on data sparsity, movie lens datasets (1M) consisting of 671 users and 163,949 product items were used by determining the preference level between 1 and 5 and user satisfaction levels of all 98,903 items being build and test the models. Methods was divided into three parts included 1) Simple Agent Module 2) Neighbor Filtering and 3) Prediction and Recommend. Simple clustering was used to create a system to provide suggestions for sparsity data. Datasets obtained from clustering represented the sample agent of dataset to being create the recommendation system. Datasets were divided into two categories, 1) Traditional Data (TD) and 2) Statistic Data (SD), and each dataset clustered by k-means clustering. The experimental results demonstrated that the number of item types in the system were recommended in the TD and Euclidean (DIS). DIS was used to find the nearest value in TD for the item list recommendation to active users in the system with the a lot of number choice of recommendation system. © 2018, ECTI Association Sirindhon International Institute of Technology. All Rights Reserved.

This research shows the ranking of university websites according to the principles of SEO, which is considered important in increasing the number of visitors. It will make more purchases and services. This is a way to rank your business website through search engines which is an important tool of digital marketing. It is similar to a university website ranking called Webometric, which represents the quality of a university website. This research has developed a system to guide the website development base on SEO by testing with samples of Rajamangala University of Technology. It has also tested with popular search engines as well as testing the use of SEO tools. Researchers find equation of website ranking based on research results. This research will serve as a guide to developing quality SEO sites and improving rankings in the SERPs, which will make the site more accessible. The purpose of website development.

The objective of this study was to introduce the recommender system based on expert and item category to match the right items to users. In this study, the expert identification was divided into 3 techniques which were 1) the experts from social network technique 2) the experts from the frequency of rating technique and 3) the experts from other user's preferences. To filter the expert users by using the frequency of rating technique and the experts from other user's preferences technique, data about item category is used. For evaluation in this study, the researcher used Epinion for the performance testing to find out errors and accuracies in the prediction process. The results of this study showed that all the presented techniques had mean absolute error score at about 0.15 and 85 percentages of accuracy, especially the expert identification combining with item category, it can reduce 60 percentages of the duration of recommendation creating.

This book includes selected contributions related to big data and data networking, presented at the 13th International Conference on Computing and Information Technology (IC2IT), which was held at the Arnoma Grand Hotel Bangkok, Thailand, July 6–7, 2017. The aim of the conference was to present emerging algorithms, methods and technologies with a high degree of originality, novelty and innovation addressing the conference theme `Mastering Data and Networking’. Section 1 and 2 discuss various aspects of data mining and corresponding applications. Section 3 focuses on speed and overhead networking optimisation problems, as well as energy problems of autonomous systems, which are becoming increasingly important. The key to addressing these problems is properly determining critical parameters. Section 4 sheds light on natural language processing, including extraction of trends and popularity and recognition of emotions as well as classic topics such as detection and classification.