• Journal of Electrical Engineering and Technology
• /
• v.13 no.2
• /
• pp.742-751
• /
• 2018

Electric vehicles (EV) are emerging as the future transportation vehicle reflecting their potential safe environmental advantages. Vehicle to Grid (V2G) system describes the hybrid system in which the EV can communicate with the utility grid and the energy flows with insignificant effect between the utility grid and the EV. The paper presents an optimal power control and energy management strategy for Plug-In Electric Vehicle (PEV) charging stations using Wind-PV-FC-Battery renewable energy sources. The energy management optimization is structured and solved using Multi-Objective Particle Swarm Optimization (MOPSO) to determine and distribute at each time step the charging power among all accessible vehicles. The Model-Based Predictive (MPC) control strategy is used to plan PEV charging energy to increase the utilization of the wind, the FC and solar energy, decrease power taken from the power grid, and fulfil the charging power requirement of all vehicles. Desired features for EV battery chargers such as the near unity power factor with negligible harmonics for the ac source, well-regulated charging current for the battery, maximum output power, high efficiency, and high reliability are fully confirmed by the proposed solution.

• Smart Structures and Systems
• /
• v.30 no.2
• /
• pp.195-207
• /
• 2022

Computational drawbacks associated with regular predictive models have motivated engineers to use hybrid techniques in dealing with complex engineering tasks like simulating the compressive strength of concrete (CSC). This study evaluates the efficiency of tree potential metaheuristic schemes, namely shuffled complex evolution (SCE), multi-verse optimizer (MVO), and beetle antennae search (BAS) for optimizing the performance of a multi-layer perceptron (MLP) system. The models are fed by the information of 1030 concrete specimens (where the amount of cement, blast furnace slag (BFS), fly ash (FA1), water, superplasticizer (SP), coarse aggregate (CA), and fine aggregate (FA2) are taken as independent factors). The results of the ensembles are compared to unreinforced MLP to examine improvements resulted from the incorporation of the SCE, MVO, and BAS. It was shown that these algorithms can considerably enhance the training and prediction accuracy of the MLP. Overall, the proposed models are capable of presenting an early, inexpensive, and reliable prediction of the CSC. Due to the higher accuracy of the BAS-based model, a predictive formula is extracted from this algorithm.

• The Journal of Information Systems
• /
• v.32 no.4
• /
• pp.229-245
• /
• 2023

Purpose This study aims to understand the inclinations of young job seekers towards "hidden champions" - small but competitive companies that are emerging as potential solutions to the growing disparity between youth-targeted job vacancies and job seekers. We utilize machine learning techniques to discern the appeal of these hidden champions. Design/methodology/approach We examined the characteristics of small and medium-sized enterprises using data sourced from the Ministry of Employment and Labor and Youth Worknet. By comparing the efficacy of five machine learning classification models (i.e., Logistic Regression, Random Forest Classifier, Gradient Boosting Classifier, LGBM Classifier, and XGB Classifier), we discovered that the predictive model utilizing the LGBM Classifier yielded the most consistent performance. Findings Our analysis of the relative significance of preference determinants revealed that industry type, geographical location, and employee count are pivotal factors influencing preference. Drawing from these insights, we propose targeted strategic interventions for policymakers, hidden champions, and young job seekers.

• Journal of Korea Artificial Intelligence Association
• /
• v.2 no.1
• /
• pp.7-14
• /
• 2024

In this paper, we explore the application of RNA-sequencing data and ensemble machine learning to predict lung cancer and treatment strategies for lung cancer, a leading cause of cancer mortality worldwide. The research utilizes Random Forest, XGBoost, and LightGBM models to analyze gene expression profiles from extensive datasets, aiming to enhance predictive accuracy for lung cancer prognosis. The methodology focuses on preprocessing RNA-seq data to standardize expression levels across samples and applying ensemble algorithms to maximize prediction stability and reduce model overfitting. Key findings indicate that ensemble models, especially XGBoost, substantially outperform traditional predictive models. Significant genetic markers such as ADGRF5 is identified as crucial for predicting lung cancer outcomes. In conclusion, ensemble learning using RNA-seq data proves highly effective in predicting lung cancer, suggesting a potential shift towards more precise and personalized treatment approaches. The results advocate for further integration of molecular and clinical data to refine diagnostic models and improve clinical outcomes, underscoring the critical role of advanced molecular diagnostics in enhancing patient survival rates and quality of life. This study lays the groundwork for future research in the application of RNA-sequencing data and ensemble machine learning techniques in clinical settings.

• Biomolecules & Therapeutics
• /
• v.17 no.1
• /
• pp.1-11
• /
• 2009

Since the Committee for Proprietary Medicinal Products (CPMP) of the European Union issued in 1997 a "points to consider" document for the assessment of the potential for QT interval prolongation by non-cardiovascular agents to predict drug-induced torsades de pointes (TdP), the QT liability has become the critical safety issue in the development of pharmaceuticals. As TdP is usually linked to delayed cardiac repolarization, international guideline (ICH S7B) has advocated the standard repolarization assays such as in vitro IKr (hERG current) and in vivo QT interval, or in vitro APD (as a follow up) as the best biomarkers for predicting the TdP risk. However, the recent increasing evidence suggests that the currently used above biomarkers and/or assays are not fully predictive for TdP, but also does not address potential new druginduced TdP due to the selective disruption of hERG protein trafficking to the cell membrane or VT and/or VF with QT shortening. There is, therefore, an urgent need for other surrogate markers or assays that can predict the proarrhythmic potential of drug candidate. In this review, we provide an ideal pre-clinical strategy to predict the potentials of QT liability and lethal arrhythmia of the drug candidates with recent issues in this field in mind, not at the expense of discarding therapeutically innovative drugs.

• International Journal of Fluid Machinery and Systems
• /
• v.7 no.2
• /
• pp.68-79
• /
• 2014

The type of turbine developed is based on the very low head of water potential source for the electric power production. The area of research is focused for the axial water turbine that can be applied at the simple site open channel with has a very low cost and environmental impact compared to the conventional hydro installation. High efficiency of axial turbine which applied to the very low potential head will made this type of turbine can be used at wider potential site. Existing irrigation weir and river area will be the perfect site for this turbine. This paper will compare the effects of the variation of swirl velocity criterion during the design of the blade of guide vane and rotor of the turbine. Effects of the swirl velocity criterion is wider known as a vortex conditions (free vortex, force vortex and swirl velocity constant), and the free vortex is the very popular condition that applied by most of turbine designer, therefore will be interesting to do a comparison against other criterion. ANSYS Fluent will be used for simulation and to determine the predictive performance obtained by each of design criteria.

• Journal of Environmental Health Sciences
• /
• v.47 no.5
• /
• pp.462-471
• /
• 2021

Background: The number of synthesized chemicals has rapidly increased over the past decade. For many chemicals, there is a lack of information on toxicity. With the current movement toward reducing animal testing, the use of toxicity big data and deep learning could be a promising tool to screen potential toxicants. Objectives: This study identified potential chemicals related to reproductive and estrogen receptor (ER)-mediated toxicities for 1135 cleaning products and 886 laundry products. Methods: We listed chemicals contained in cleaning and laundry products from a publicly available database. Then, chemicals that potentially exhibited reproductive and ER-mediated toxicities were identified using the European Union Classification, Labeling and Packaging classification and ToxCast database, respectively. For chemicals absent from the ToxCast database, ER activity was predicted using deep learning models. Results: Among the 783 listed chemicals, there were 53 with potential reproductive toxicity and 310 with potential ER-mediated toxicity. Among the 473 chemicals not tested with ToxCast assays, deep learning models indicated that 42 chemicals exhibited ER-mediated toxicity. A total of 13 chemicals were identified as causing reproductive toxicity by reacting with the ER. Conclusions: We demonstrated a screening method to identify potential chemicals related to reproductive and ER-mediated toxicities utilizing chemical toxicity big data and deep learning. Integrating toxicity data from in vivo, in vitro, and deep learning models may contribute to screening chemicals in consumer products.

• Korean Journal of Radiology
• /
• v.22 no.7
• /
• pp.1163-1171
• /
• 2021

Objective: To analyze the correlations between intraoperative ultrasound and MRI metrics of the spinal cord in degenerative cervical myelopathy and identify novel potential predictive ultrasonic indicators of neurological recovery for degenerative cervical myelopathy. Materials and Methods: Twenty-two patients who underwent French-door laminoplasty for multilevel degenerative cervical myelopathy were followed up for 12 months. The Japanese Orthopedic Association (JOA) scores were assessed preoperatively and 12 months postoperatively. Maximum spinal cord compression and compression rates were measured and calculated using both intraoperative ultrasound imaging and preoperative T2-weight (T2W) MRI. Signal change rates of the spinal cord on preoperative T2W MRI and gray value ratios of dorsal and ventral spinal cord hyperechogenicity on intraoperative ultrasound imaging were measured and calculated. Correlations between intraoperative ultrasound metrics, MRI metrics, and the recovery rate JOA scores were analyzed using Spearman correlation analysis. Results: The postoperative JOA scores improved significantly, with a mean recovery rate of 65.0 ± 20.3% (p < 0.001). No significant correlations were found between the operative ultrasound metrics and MRI metrics. The gray value ratios of the spinal cord hyperechogenicity was negatively correlated with the recovery rate of JOA scores (ρ = -0.638, p = 0.001), while the ventral and dorsal gray value ratios of spinal cord hyperechogenicity were negatively correlated with the recovery rate of JOA-motor scores (ρ = -0.582, p = 0.004) and JOA-sensory scores (ρ = -0.452, p = 0.035), respectively. The dorsal gray value ratio was significantly higher than the ventral gray value ratio (p < 0.001), while the recovery rate of JOA-motor scores was better than that of JOA-sensory scores at 12 months post-surgery (p = 0.028). Conclusion: For degenerative cervical myelopathy, the correlations between intraoperative ultrasound and preoperative T2W MRI metrics were not significant. Gray value ratios of the spinal cord hyperechogenicity and dorsal and ventral spinal cord hyperechogenicity were significantly correlated with neurological recovery at 12 months postoperatively.

• Journal of Ecology and Environment
• /
• v.38 no.4
• /
• pp.529-536
• /
• 2015

This study was performed to provide important basic data for the preservation and management of Scopura laminata, a species endemic to Korea, by elucidating the spatial characteristics of its present, potential, and future distribution areas. Currently, this species is found in the Odaesan National Park area of South Korea and has been known to be restricted in its habitat due to its poor mobility, as even fully grown insects do not have wings. Utilizing the MaxEnt model, 20 collection points around Odaesan National Park were assessed to analyze and predict spatial distribution characteristics. The precision of the MaxEnt model was excellent, with an AUC value of 0.833. Variables affecting the potential distribution area of S. laminata by more than 10% included the range of annual temperature, seasonality of precipitation, and precipitation of the driest quarter, in order of greatest to least impact. Compared to the current potential distribution area, no significant difference in the overall habitable area was predicted for the 2050s or 2070s. It was, however, demonstrated that the potential habitable area would be reduced in the 2050s by up to 270.3 km from the current area of 403.9 km; further, no potential habitable area was anticipated by the 2070s according to our predictive model. Taken together, it is anticipated that this endemic species could be significantly affected by climate changes, and hence effective countermeasures are strongly warranted for the preservation of habitats and species management.

• Journal of Environmental Science International
• /
• v.18 no.2
• /
• pp.129-139
• /
• 2009

Development of an artificial neural network model was presented to predict the daily maximum $SO_2$ concentration in the urban-industrial area of Ulsan. The network model was trained during April through September for 2000-2005 using $SO_2$ potential parameters estimated from meteorological and air quality data which are closely related to daily maximum $SO_2$ concentrations. Meteorological data were obtained from regional modeling results, upper air soundings and surface field measurements and were then used to create the $SO_2$ potential parameters such as synoptic conditions, mixing heights, atmospheric stabilities, and surface conditions. In particular, two-stage clustering techniques were used to identify potential index representing major synoptic conditions associated with high $SO_2$ concentration. Two neural network models were developed and tested in different conditions for prediction: the first model was set up to predict daily maximum $SO_2$ at 5 PM on the previous day, and the second was 10 AM for a given forecast day using an additional potential factors related with urban emissions in the early morning. The results showed that the developed models can predict the daily maximum $SO_2$ concentrations with good simulation accuracy of 87% and 96% for the first and second model. respectively, but the limitation of predictive capability was found at a higher or lower concentrations. The increased accuracy for the second model demonstrates that improvements can be made by utilizing more recent air quality data for initialization of the model.