‪Chatklaw Jareanpon

Dr. ‪Chatklaw Jareanpon

computer science, Mahasarakham University, Maha Sarakham, Thailand
orcid

Research Focus & Metrics

Academic Career Length

23Years (2004-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 January 2004
Verified
Conference Paper
Closed access
‪Chatklaw Jareanpon‪Chatklaw Jareanpon
‪Chatklaw Jareanpon
WPW. Pensuwon
W. Pensuwon
RFR.J. Frank
R.J. Frank
NDN. Davey
N. Davey
Rainfall prediction is a challenging task, especially in a modern world facing the major environmental problem of global warming. The proposed method uses an adaptive radial basis function neural network mode with a specially designed genetic algorithm (GA) to obtain the optimal model parameters. A significant feature of the adaptive radial basis function network is that it is able create new hidden units and solve the spread factor problem using a genetic algorithm. It is shown that the evolved parameter values improved performance.

28 April 2022 in CURRENT APPLIED SCIENCE AND TECHNOLOGY
Verified
Journal Article
Closed access
PDF file
WNWongpanya Nuankaew
Wongpanya Nuankaew
PNPratya Nuankaew
Pratya Nuankaew
DDDamrongdet Doenribram
Damrongdet Doenribram
‪Chatklaw Jareanpon‪Chatklaw Jareanpon
‪Chatklaw Jareanpon
The Twitter platform is a popular tool that is widely used by researchers to collect data on users’ personal lives, feelings and emotions. These data sets can be further analyzed using text mining techniques to predict the disorder of depression. There are nine symptoms of depression that are classified by American Psychiatric Association using DSM-5 criteria. The symptoms can be difficult to identify effectively. The unweighted vote ensemble is not practical for multi-class data. Therefore, this research proposes the multi-objective optimization algorithms for depressive symptom prediction modeling (MOADSP) for the weighted voting ensemble, which can improve its effectiveness compared to the singer model. The objectives of this research were 1) to find the appropriate number of features; 2) to improve the weights of the prediction models based on the recall of the class for the ensemble; and 3) to compare the performance of the single, unweighted, and weighted voting ensemble models for depressive disorder. An information gain was used to select the features. The single classification techniques used in the experiment that had their frameworks tested were the Naïve Bayes, Random Forest, and K-Nearest techniques, while the vote ensemble models used were the unweighted and weighted models. MOADSP was applied to the weighted vote ensemble models. The results showed that the best recall classifier was KNN (98.60%), and the highest recall classifier was AVG TP weighted (98.43%) for the training model. The highest recall in the class depressive classifier was AVG TP weighted (80.00%) for the testing. This proposed method was beneficial for the prediction of depressive disorder.

30 September 2023 in Journal of Education and Community Health
Verified
Journal Article
Open access
PDF file
AMAravan Mungvongsa
Aravan Mungvongsa
SYSuneerat Yangyuen
Suneerat Yangyuen
‪Chatklaw Jareanpon‪Chatklaw Jareanpon
‪Chatklaw Jareanpon
TSThidarat Somdee
Thidarat Somdee
Background: Health literacy (HL) is an indicator of health outcomes, but its role in dietary intake has received little attention. Excessive dietary intake increases the risk of metabolic syndrome (MetS). Therefore, this study aimed to investigate the HL score, dietary intake, and nutrient intake of participants and the relationship between HL score and dietary intake among adults with MetS in Thailand. Methods: In this cross-sectional study, 2527 adults aged 18–59 years in primary care services, Phetchaburi, Thailand were included in the study using a multistage sampling technique. We determined HL scores using the Health Literacy Questionnaire (HLQ) and dietary intake using a semi-quantitative Food Frequency Questionnaire. We used multiple linear regression analysis to investigate the associations between HL score and dietary intake. Results: HL scores were significantly lower in patients with MetS compared with participants without it (P<0.05). Participants with MetS had significantly higher calorie and fat intake than participants without it (P<0.05), and participants with MetS had higher fat and lower carbohydrate intake. The results of multiple linear regression showed a significant negative association between HL score and dietary intake, after controlling for potential confounding variables (ꞵ=−0.053, P<0.05). Conclusion: Our findings suggest that low HL score is associated with high dietary intake. Therefore, improving HL might play an important role in reducing dietary intake to decrease the risk of MetS.

29 March 2024 in ECTI Transactions on Computer and Information Technology (ECTI-CIT)
Verified
Journal Article
Open access
PDF file
ATApirak Tooltham
Apirak Tooltham
Suchart KhummaneeSuchart Khummanee
Suchart Khummanee
‪Chatklaw Jareanpon‪Chatklaw Jareanpon
‪Chatklaw Jareanpon
MNMontree Nonphayom
Montree Nonphayom
The agricultural sector is projected to need more labor as a result of declining interest in careers within this domain. Despite the escalating demand for agricultural goods, previous endeavors to mitigate this challenge through the deployment of robotic prototypes have encountered hindrances such as issues pertaining to automation, adaptability to varying tasks, and the financial burdens associated with development. To address this exigency, we have developed an automated cultivation robot utilizing advancements in the Internet of Things (IoT), Image Processing, and artificial Intelligence (AI) for seeding in pots. The robot demonstrates the capacity to sow seeds in 257 pots per hour, accomplish a mission within 12.53 minutes, traverse at a velocity of 360 meters per hour, and seed pots at a rate of 13.37 seconds per pot. It possesses an operational duration of approximately two hours, completing nine cycles and seeding 486 pots on a single charge. Notably, the robot exhibits a mission success rate of 1.00 and a seeding accuracy 0.78. Moreover, it features an adaptable workspace and a lightweight frame weighing 20 kg, rendering it a cost-effective solution for mass production.

27 August 2026 in Applied Sciences
Verified
Journal Article
Open access
PDF file
‪Chatklaw Jareanpon‪Chatklaw Jareanpon
‪Chatklaw Jareanpon
KKKhanabhorn Kawattikul
Khanabhorn Kawattikul
Joint-Embedding Predictive Architectures (JEPAs) have recently emerged as a promising paradigm for self-supervised representation learning by predicting latent embeddings from partial observations rather than reconstructing raw inputs. Although JEPA has demonstrated considerable success in computer vision and large-scale language models, the design of lightweight JEPA-inspired representation learning frameworks for natural language remains insufficiently understood. This paper proposes a lightweight span-masked JEPA-inspired framework that formulates language representation learning as latent semantic prediction from partially observed contextual inputs. The framework predicts target embeddings generated by a momentum-updated encoder while avoiding direct token reconstruction, thereby emphasizing semantic abstraction over lexical recovery. To investigate the influence of objective function design, cosine similarity and Variance–Invariance–Covariance Regularization (VICReg) are systematically compared under identical architectural and training conditions, with and without auxiliary masked language modeling (MLM) supervision. The proposed framework is evaluated on one controlled dataset and six benchmark datasets, including AG News, SST-2, IMDb, TREC, DBPedia, and Yelp Polarity. Experimental results consistently demonstrate that the cosine-based JEPA objective without auxiliary MLM supervision provides the best balance between representation quality, computational efficiency, and downstream linear probing performance. Statistical significance analysis using the Friedman test and pairwise Wilcoxon signed-rank tests with Holm correction further confirms that the cosine-based objective significantly outperforms the remaining configurations, while no significant difference is observed between VICReg with and without MLM supervision. The findings indicate that latent semantic prediction alone is sufficient for learning robust language representations under the proposed lightweight framework, whereas auxiliary token-level reconstruction does not consistently improve downstream performance despite increasing computational cost. Overall, this work provides reproducible empirical evidence and practical guidance for objective function design in lightweight JEPA-inspired language representation learning, offering an efficient experimental foundation for future self-supervised language models.

22 October 2026 in Symmetry
Verified
Journal Article
Open access
PDF file
Suchart KhummaneeSuchart Khummanee
Suchart Khummanee
Olarik SurintaOlarik Surinta
Olarik Surinta
‪Chatklaw Jareanpon‪Chatklaw Jareanpon
‪Chatklaw Jareanpon
Coordinating whole-body shape generation, geometric object caging, and collision-aware transportation remains challenging for hyper-redundant robots. This study presents a framework integrating Transformer-based backbone prediction, cage-preserving geometric refinement, and neural shape control for a planar 50-link robot operating among static and moving obstacles. The framework was evaluated in five simulated scenes under idealized kinematic assumptions without explicit actuator dynamics, contact forces, or sensing uncertainty. The full implementation achieved a geometric cage success rate of 100%, with mean cage-center and cage-radial errors of 0.001786 and 0.000062, respectively. However, the mean robot–obstacle and object–obstacle collision rates were 75.22% and 15.87%, and none of the five scenes satisfied the complete task-success criterion. A controlled ablation of soft cage attraction produced nearly unchanged performance when hard cage projection was retained, indicating limited incremental benefit from the soft attraction term under the evaluated conditions. This experiment does not isolate the contribution of symmetry preservation as a whole. A dedicated CPU benchmark measured a mean control-cycle latency of 128.792 ± 2.401 ms, with shape optimization accounting for approximately 51.4% of the mean cycle time. The findings demonstrate consistent maintenance of local enclosure geometry but do not establish collision-free transportation or physically feasible manipulation. Further development should prioritize whole-body feasibility, collision-aware motion planning, and validation across additional training seeds and physically realistic environments.