Nowadays chatbots have been widely adopted in many industries to automatically answer users' questions and requests via chat interfaces. While it has become much easier develop a chatbot system, the system itself is complex nature. It challenge evaluate compare various systems terms of effectiveness, efficiency, goal achievability, ability satisfy users. This paper presents survey, starting from literature review, architecture, evaluation methods/criteria, comparison methods. Focused on...

Urban flooding is a major problem in Thailand. An essential countermeasure towards better management to forecast flood water levels the real-time manner. Most existing early warning systems (EWS) Thailand contain lot of miscalculations when they face with real situations. Towards prediction improvement, this paper presents hydrological modeling augmented alternative five machine learning techniques; linear regression, neural network Bayesian regression and boosted decision tree regression....

Computer vision syndrome (CVS) describes a group of eye and vision-related problems that result from prolonged digital device use. This study aims to assess the prevalence associated factors CVS among students during lockdown resulting COVID-19 pandemic.A cross-sectional, online, questionnaire-based performed high school in Thailand.A total 2476 students, with mean age 15.52±1.66 years, were included this study. The number hours use per day (10.53±2.99) increased pandemic compared before its...

Predicting protein subcellular location is one of major challenges in Bioinformatics area since such knowledge helps us understand functions and enables to select the targeted proteins during drug discovery process. While many computational techniques have been proposed improve predictive performance for location, they several shortcomings. In this work, we propose a method solve three main issues techniques; i) manipulation multiplex which may exist or move between multiple cellular...

Abstracts Background The aim of this study was to evaluate the most effective combination autoregressive integrated moving average (ARIMA), a time series model, and association rule mining (ARM) techniques identify meaningful prognostic factors predict number cases for efficient COVID-19 crisis management. Methods 3685 patients admitted at Thailand’s first university field hospital following four waves infections from March 2020 August 2021 were analyzed using its derivative exogenous...

Promoting reflective thinking is becoming increasingly important in helping learners develop strategies to apply new information unpredictable situations during their daily activities. Reflective writing, based on Kolb’s experiential learning cycle, could be one method for promoting that allows consider experiences and transform them into transferable knowledge, which can applied contexts. However, cannot sufficiently practice writing; thus, they learn from experiences. Therefore, our...

Purpose: The purpose of this study was to evaluate the diagnostic performance deep learning (DL) anterior segment optical coherence tomography (AS-OCT) as a plateau iris prediction model. Design: We used cross-sectional development and validation DL system. Methods: conducted collaboration between referral eye center an informative technology department. enrolled 179 eyes from 142 patients with primary angle closure disease (PACD). All had remaining appositional after iridotomy. Each scanned...

We present LLaVAC, a method for constructing classifier multimodal sentiment analysis. This leverages fine-tuning of the Large Language and Vision Assistant (LLaVA) to predict labels across both image text modalities. Our approach involves designing structured prompt that incorporates unimodal fine-tune LLaVA, enabling it perform classification effectively. Experiments on MVSA-Single dataset demonstrate LLaVAC outperforms existing methods in analysis three data processing procedures. The...

To add more value on YouTube, a popular portal of social media clips, it is worth recognizing automatically the mood clip using comments given to such clip. This paper presents method classify emotion Thai YouTube Six basic emotions considered are Anger, Disgust, Fear, Happiness, Sadness and Surprise. Performances three alternative machine learning algorithms, namely multinomial naïve Bayes (MNB), decision tree (DT) support vector (SVM) compared. As result, SVM achieves highest accuracy in...

Taxis play an important role in urban transportation. Understanding the taxi demand future gives opportunity to organize fleet better. It also reduces waiting time of passengers and cruising drivers. Even, there are some works proposed predict but few studies that consider function areas such as hospital area, department store residential tourist attraction. One predictive model may not fit with all types area. We use a point interest (POI) match place study area different function. In this...

Road traffic mortalities (RTMs), a leading cause of death globally, mostly occur in low- and middle-income countries, having sufficient healthcare resources could lower the number these fatalities. Our study aimed to illustrate incidence RTMs per 100,000 population compare distribution from 2011 2021 with rates 77 provinces Thailand. We divided into adults (≥ 15 years) children (0-14 years). Lorenz curve Gini coefficient were used measure level equality hospital relation across country. The...

Cloud data warehouse (CDW) platforms have been offered by many cloud service providers to provide abundant storage and unlimited accessibility business users. Sensitive (DW) consisting of dimension fact is typically encrypted before it outsourced the cloud. However, query over DW not practically supported any analytical tools. The Searchable Encryption (SE) technique palpable for supporting keyword searches data. Although SE schemes introduced their own unique searching methods based on...

The imbalanced dataset problem triggers degradation of classification performance in several data mining applications including pattern recognition, text categorization, and information filtering tasks. To improve emotion performance, we use a sampling-based algorithm called SMOTE, which oversamples instances minority class to the number those from majority class. YouTube was balanced using SMOTE technique tested three machine learning algorithms, namely multinomial Naïve Bayes (MNB),...

Many megacities in the world, especially Bangkok, are facing severe congestion road traffic and public transport, particularly during peak hours. This situation (a) worsens quality of life, (b) releases emissions causing air pollution climate change, (c) subsequently creates requests for massive investment transport infrastructure, which easily exceed budget’s limit. Instead solving problem by supply-side strategies, applying ICT-based solutions to reform people’s daily activities, commuting...

Generalized association rule mining is an extension of traditional to discover more informative rules, given a taxonomy. We describe formal framework for the problem generalized rules. In framework, The subset-superset and parent-child relationships among itemsets are introduced present different views itemsets, i.e. lattice taxonomies k-generalized respectively. optimization technique reduce time consumed by applying two constraints each which corresponds view itemsets. process, new set...

Renewing feeder transit services is crucial for maximizing the potential of public transport system and advancing toward a decarbonized Bangkok. Herein, we propose smart service using small electric vehicles limited areas (sois) target facilities in central A 1-year proof concept was conducted to verify feasibility effectiveness this assess its deployment as new mobility service. We found that 1) introduction sois can considerably reduce both road congestion levels CO2 emissions owing...

Some languages including Thai, Japanese and Chinese do not have explicit word boundary. This causes the problem of boundary ambiguity that results in decreasing accuracy information retrieval. paper proposes a new technique so-called character clustering to reduce Thai documents hence improve searching efficiency. To investigate efficiency, set experiments using newspapers is conducted both non-indexing indexing approaches. The experimental show our method outperform traditional methods...

For languages without word boundary delimiters, dictionaries are needed for segmenting running texts. This figure makes segmentation accuracy depend significantly on the quality of dictionary used analysis. If is not sufficiently good, it will lead to a great number unknown or unrecognized words. These words certainly reduce accuracy. To solve such problem, we propose method based decision tree models. Without use dictionary, specific information, called syntactic attribute, applied identify...

Most previous works on text categorization applied term occurrence frequency and inverse document for representing importance of terms. This work presents an analysis class in centroid-based categorization. There are two aims this paper. The first one is to find appropriate functions frequency. other the key factors using experimental results show that factors, which improve classification accuracy, numbers few-class terms most-class When large obtained, logarithmic function most effective...