KMITL

Permanent URI for this communityhttps://dspace.kmitl.ac.th/handle/123456789/1

Browse

Search Results

Now showing 1 - 10 of 24

Some of the metrics are blocked by your
consent settings
A Comparative Study of Deep Reinforcement Learning Agents for Gold Trading with Technical Indicators and LLM-Filtered News Sentiment
(2026-06-16)
Thanasarn, Thanapong
;
Anuntachai, Anuntapat
;
Netisopakul, Ponrudee
Integrating macroeconomic news into deep reinforcement learning (DRL) for daily gold (XAU/USD) trading remains challenging. This study implements a natural language processing pipeline using majority voting from three large language models (Llama-3.1-8B-Instruct, Qwen3-8B, and Gemma-3-12B-it) to filter articles from The New York Times from 2010 to 2024, yielding 592 relevant articles that are subsequently assigned sentiment scores using FinBERT. We evaluate A2C, PPO, and SAC agents in a FinRL and Stable-Baselines3 framework using a four-fold walk-forward expanding-window protocol and Optuna hyperparameter tuning. The models are compared under two feature settings: (1) six technical indicators only and (2) technical indicators combined with sentiment features. Results show that incorporating LLM-filtered sentiment can modestly improve trading performance for some agents. PPO with sentiment achieves the best average cumulative return (10.05%) and Sharpe ratio (0.74), compared with its indicator-only version (9.80%, 0.72), and slightly outperforms the Buy-and-Hold (B&H) baseline (9.58%, 0.71). A2C also improves with sentiment (9.50% to 9.90%), while SAC shows no improvement in this setting. These findings suggest that LLM-filtered sentiment provides a modest benefit for some DRL agents in daily gold trading in our experiments.
Some of the metrics are blocked by your
consent settings
Personalized Health Insurance Recommendation with Retrieval-Augmented Generation: A Study of Lexical, Dense, and Hybrid Retrieval
(2026-06-16)
Rujireksareekul, Phakon
;
Anuntachai, Anuntapat
;
Netisopakul, Ponrudee
Health insurance information in Thailand is available on many company websites and is presented in different formats, making it difficult for users to compare coverage, conditions, and benefits. In this study, we propose a personalized health insurance recommendation system using Retrieval-Augmented Generation (RAG). The system retrieves information using three methods: BM25, dense retrieval, and hybrid retrieval with Reciprocal Rank Fusion. The language model analyzes the retrieved insurance plans and recommends suitable options based on the user query. Experimental results show that dense retrieval provides the best overall performance, while hybrid retrieval performs better than lexical search. The proposed RAG system also maintains practical response latency, making it suitable for interactive applications.
Some of the metrics are blocked by your
consent settings
Machine Learning Models for Multi-Horizon Classification of Bitcoin Future Price Movements
(2026-06-16)
Boonpai, Sirawat
;
Netisopakul, Ponrudee
;
Anuntachai, Anuntapat
This study presents a comprehensive framework for multi-horizon classification of Bitcoin futures price movements using machine learning. Five models of Logistic Regression, Random Forest, XGBoost, LightGBM, and CatBoost were systematically evaluated across three timeframes (4h, 12h, 1d) and four classification schemes: binary (up/down) and multi-class (up/down/stable) with thresholds of 0.5%, 1.0%, and 1.5%, totaling 60 distinct experimental configurations. Rather than pursuing a strong predictive performance, this study prioritizes a rigorous comparative analysis across multiple dimensions to identify which combinations of model, timeframe, and classification scheme are most effective for Bitcoin futures. The results demonstrate that binary classification achieves the best predictive performance, with shorter timeframes yielding significantly better results, confirming the effectiveness of technical indicators in capturing the rapid price dynamics of Bitcoin futures. Notably, CatBoost achieved the highest F1-score for binary classification, while Random Forest proved the most robust model across diverse configurations. Feature importance analysis revealed that momentum-based indicators are the dominant predictors of price direction, while volatility features play a critical role in distinguishing sideways movements from directional ones in multi-class settings. Furthermore, the study demonstrates that narrower classification thresholds (0.5%) introduce noisier class boundaries and degrade performance, whereas wider thresholds (1.0%-1.5%) yield more stable results. These findings provide actionable guidelines for algorithmic trading in Bitcoin futures and establish a reproducible benchmark for future research.
Some of the metrics are blocked by your
consent settings
Virtue-Based Thai Folktale Recommendation with Ensemble LLMs
(2026-06-16)
Daeng-Am, Wassana
;
Anuntachai, Anuntapat
;
Netisopakul, Ponrudee
Moral learning plays an important role in early childhood education, yet teachers often spend considerable time identifying moral lessons in stories and deciding which virtues they represent. This study investigates whether a multi-model ensemble strategy can produce more teacher-aligned moral extractions from Thai folktales than a single-LLM baseline. We propose an LLM-based ensemble framework that extracts concise moral statements and classifies them into eight core virtues promoted by the Thai Ministry of Education: diligence, frugality, honesty, discipline, politeness, cleanliness, unity, and kindness. The framework combines outputs from Google Gemini 2.0 Flash, OpenAI GPT-4o-mini, and Anthropic Claude 3.5 Sonnet through a semantic consensus mechanism using BGE-M3 embeddings and majority voting. The system is evaluated on a corpus of 200 Thai folktales annotated by three experienced early childhood educators using multi-label metrics including Hamming Loss, Jaccard Similarity, and Exact Match. The results show that the ensemble approach achieves a Hamming Loss of 0.208, Jaccard Similarity of 0.564, and Exact Match of 0.175, consistently outperforming all single-model baselines. These findings suggest that consensus-driven ensemble inference provides a more robust and teacher-aligned foundation for automated moral education tools in Thai NLP.
Some of the metrics are blocked by your
consent settings
Enhancing Industrial Automation: Real-Time Detection of Plastic Spoon Presence in Cup Filling Using Deep Learning and Plc Integration
(2025-01-01)
Chawlert, Kanoksak
;
Anuntachai, Anuntapat
This paper presents a real-time detection system to enhance industrial automation in cup-filling processes by identifying the presence of plastic spoons using Deep Learning and Programmable Logic Controller (PLC) integration. The proposed system employs the YOLO (You Only Look Once) object detection model to process images captured by a vision camera. The model runs on a PC and communicates with a BECKHOFF PLC to perform automated decisionmaking based on detection results. A hardware simulation environment was developed, consisting of a servo-driven rotary table and camera setup, mimicking an actual production line. The system detects two classes: cups with spoons (acceptable) and cups without spoons (defective). Upon detection, the result is sent to the PLC, which initiates appropriate control actions. Experimental testing demonstrates that the proposed AI-integrated automation system improves detection accuracy, reduces errors from conventional sensors, and ensures better product quality control in real-time operations.
Some of the metrics are blocked by your
consent settings
Egg Defect Detection and Classification in Boiled Egg Industry with Surface Disturbance Removal on the Eggshell Based on Image Processing
(2025-01-01)
Chotchawalkul, Sasikan
;
Chaipanya, Oraya
;
Anuntachai, Anuntapat
In the boiled egg industry, quality inspection is typically conducted twice: before eggs are transported into the conveyor-based boiling system (before boiling), and after they exit the water-based cooling system prior to packaging (after cooling). These inspections are commonly carried out through human visual assessment, which demands substantial human resources and time. This paper presents an automated system for detecting and classifying defective eggs-such as those with cracks, dents, rough shells, and other surface anomalies-using image processing techniques. The system is designed to enhance the visibility of such defects while minimizing the impact of production-related surface disturbances, including water stains, reflections from the cooling process, and red stamps from imported eggs. The proposed system comprises two main approaches: (1) Defects Detection Method, which classifies eggs into two categories: intact and defective; and (2) Pixel Counting and Comparison Method, which classifies eggs into three categories: intact, cracked or dented, and exploded eggs. This system offers a practical and efficient solution for the egg processing industry, reducing reliance on human labor, minimizing inspection time, and lowering hardware requirements for industrial implementation.
Some of the metrics are blocked by your
consent settings
Skill Level Recognition in Writing for Elementary School Students in Thailand Using Image Processing Technology
(2024-01-01)
Jarusitratti, Nattwat
;
Laohakul, Krittatee
;
Anuntachai, Anuntapat
In the current context, the problem of developmental writing difficulties in children is of great importance for school-age children. Diagnosing whether a child has developmental writing difficulties requires the use of writing skills assessments. These assessments are used by professionals to evaluate and diagnose any abnormalities in a child's writing development. However, there are limitations in terms of format, as they often rely on expert physicians for diagnosis. This creates a significant need for human resources. To address this, we have designed a method for scoring based on writing skills assessments, utilizing image processing technology and criteria from existing standards. The scoring criteria include three aspects: article writing position, article format, and copying speed. For article writing position, we find the centroid of the text. Article format is assessed based on the aesthetics of the written article, which should form a parallelogram. Lastly, copying speed is determined by the number of lines visible on the paper, using pixel frequency analysis.
Some of the metrics are blocked by your
consent settings
System for Analysis and Verification of Exercise Postures with Equipment
(2024-01-01)
Kamcharoen, Chayanee
;
Boriboon, Pragasit
;
Anuntachai, Anuntapat
This paper is initiated in response to affectation from pandemic. The people demand to improve their heath by themselves, and the exercise equipment is easier to install at home. Exercising with equipment requires fundamental knowledge to avoid any injuries and reduce ineffective performance. The system analyzes exercise postures using equipment, encompassing 3 poses: Deadlift, Lat Pull Down, and Bench Press. Analysis is comparing the alignment of skeletal joints in the body from exercise videos through image processing and comparing correctness against expert's movement. Results and recommendations are displayed on a web application which is developed by the Django Framework. Test results indicate an improvement in users' exercise direction tendencies.
Some of the metrics are blocked by your
consent settings
Lung Cancer Prediction Model from Chest X-Ray Images
(2024-01-01)
Chaiyathed, Chayodom
;
Thanesmaneekul, Ekawit
;
Anuntachai, Anuntapat
Lung cancer is one of the leading causes of death globally. Early diagnosis of lung cancer is crucial for treatment and prognosis. Traditional medical techniques, such as chest x-rays, have limitations in the early diagnosis of lung cancer. This paper develops an image classification model for chest CT scans using deep learning with transfer learning techniques. The data is divided into three parts: a training set, a testing set, and a validation set. The development of this model can be applied to improve the efficiency of early lung cancer diagnosis, reduce the risk of human errors, and increase workflow efficiency in hospitals. In this paper, a model is developed to distinguish between normal images and images with lung cancer. This model can potentially assist physicians in accurately and rapidly diagnosing lung cancer.
Some of the metrics are blocked by your
consent settings
Spacing and Stressing Extracting of Sentences System
(2024-01-01)
Ngodngamjaras, Peerach
;
Teethawatthanakorn, Napatrapee
;
Anuntachai, Anuntapat
Word stressing and spacing to convey the meaning is an important part of speaking English. Audiences are more likely to engage with a speaker who uses proper word stress and spacing. Stress can be observed from louder pronunciation, and higher frequency. Space can be observed from breathing and the length of time before starting a new sentence. These things, if those who would like to practice English are inexperienced or do not listen to native speakers often, they may not be able to remember the points of stressing and spacing correctly. As a result, speech does not flow smoothly. Nowadays, there are no tutorials or visualizations that clearly point out where the spacing is. For this reason, the organizer decided to create a program that would identify word stress and spacing to display on a dashboard that would be developed into a website in the future.

KMITL

Browse

Filters

Settings

Sort By

Results per page

Search Results