Galaxy evolution studies require the measurement of the physical properties of galaxies at different redshifts. In this work, we build supervised machine learning models to predict the redshift and physical properties (gas-phase metallicity, stellar mass, and star formation rate) of star-forming galaxies from the broad-band and medium-band photometry covering optical to near-infrared wavelengths, and present an evaluation of the model performance. Using 55 magnitudes and colors as input features, the optimized model can predict the galaxy redshift with an accuracy of σ(Δz/1+z) = 0.008 for a redshift range of z < 0.4. The gas-phase metallicity [12 + log(O/H)], stellar mass [log(Mstar)], and star formation rate [log(SFR)] can be predicted with the accuracies of σNMAD = 0.081, 0.068, and 0.19 dex, respectively. When magnitude errors are included, the scatter in the predicted values increases, and the range of predicted values decreases, leading to biased predictions. Near-infrared magnitudes and colors (H, K, and H −K), along with optical colors in the blue wavelengths (m425–m450), are found to play important roles in the parameter prediction. Additionally, the number of input features is critical for ensuring good performance of the machine learning model. These results align with the underlying scaling relations between physical parameters for star-forming galaxies, demonstrating the potential of using medium-band surveys to study galaxy scaling relations with large sample of galaxies.

Background: The increasing demand for real-time professional fitness coaching has led to a need for accurate exercise posture recognition using artificial intelligence. Objectives: To compare the performance of Feedforward Neural Network (FNN) and Stacked Long Short-Term Memory (LSTM) models in classifying fitness posture images using detailed joint coordinate labeling. Design: Comparative analysis of machine learning models using a labeled dataset of fitness posture images. Methods: A dataset from AI-hub containing images and data of 41 exercises was used. Five exercises were selected and processed using a custom program. Data was converted from JSON to CSV format, augmented with joint condition information, and analyzed using Google Colab. Results: The best FNN model achieved a training error of 1.21% and test error of 9.08%. The Stacked LSTM model demonstrated superior performance with a training error of 1.05% and test error of 6.09%. Conclusion: Both FNN and Stacked LSTM models effectively classified sequential fitness images, with Stacked LSTM showing superior performance. This indicates the potential of Stacked LSTM models for accurate fitness posture classification in real-time coaching scenarios.

Background: Artificial intelligence (AI) research on physical fitness posture estimation has been limited by a lack of comprehensive datasets and guidelines. This study analyzes the fitness image dataset provided by Korea's AIHub platform to advance posture estimation algorithms from exercise prescription and behavioral analysis perspectives. Objectives: To analyze fitness movements and guide correct exercise posture using AI-based visual and auditory feedback. Design: Descriptive analysis of a large-scale dataset. Methods: The study examined image and JSON labeling files from AI-Hub, analyzing 6.39 million fitness images across 41 exercise types. Data structure, exercise states, and annotation characteristics were analyzed in detail. Results: The dataset encompasses 816 distinct exercise states, captured from five camera angles with 24 key body points labeled per posture. Exercises were categorized into full-body workouts (17), barbell/dumbbell exercises (16), and furniture exercises (8). Gender distribution was 76% male and 24% female, with 41% in the 27-29 age group. The dataset allows for detailed analysis of correct and incorrect postures. Conclusion: This comprehensive analysis of the AI-Hub fitness dataset provides a robust foundation for developing AI models for fitness posture evaluation and feedback, benefiting exercise coach web/app service developers.

We previously reported that pyridoxine and its derivatives exert antidiabetic effects by alleviating postprandial hyperglycemia via inhibition of carbohydrate-hydrolyzing enzymes in normal sprague–dawley (SD) rats. In this study, we aimed to further evaluate whether long-term pyridoxal supplementation decreases the blood glucose levels using SD rats. SD rats were randomly assigned to groups fed a high-carbohydrate diet (66.1% cornstarch) with or without pyridoxal (4%) for 36 days. Changes in body weight, blood glucose levels, and food intake were measured daily for 36 days. Dietary supplementation with pyridoxal significantly decreased the blood glucose levels (P<0.001) and body weight (P<0.001) in mice. Glycated hemoglobin (HbA1c) levels, which are good indicators of plasma glucose concentrations over prolonged periods, were also significantly decreased over five weeks (P<0.001). Similarly, dietary treatment with Acarbose ® (0.04%), a positive control, also significantly decreased the blood glucose and HbA1c levels and body weight. Overall, our findings suggest that pyridoxal inhibits weight gain and alleviates postprandial hyperglycemia by decreasing glucose absorption and HbA1c levels.

진저 비어는 생강과 설탕을 이용한 발효 음료로, 가볍 게 탄산화된 특유의 매운맛이 특징이며, 주로 가정에서 직 접 만들어진다. Ginger bug라는 스타터 컬처를 사용한 자 연적 발효 과정을 통해 만들어지며, 이는 상업용 음료와 달리 발효된 상태의 살아있는 미생물을 포함한다. 이 연 구는 두 가지 다른 방법으로 가정용 진저 비어를 직접 제 조하여 진저 비어의 미생물 군집의 변화를 분석하고자 하 였다. 레시피 1과 2의 발효 결과, 총 균수(aerobic plate count, APC)는 최대 6 log CFU/mL에 도달했고, 효모와 곰팡이 수(yeast and mold, YM)는 6.5 log CFU/mL로 가 장 높았다. 레시피 2에서는 진저 비어를 만들기 전에 ginger bug를 발효하였으므로 알코올 함량이 0.655%까지 증가한 반면, 레시피 1에서는 0.15% 미만이었다. 다양성 분석 결 과, ginger bug에서 높은 수준의 Enterobacteriaceae가 발견 되어 발효 과정과 재료 취급이 미생물 군집 변화에 영향 을 미쳤음을 시사했다. 생강과 진저 비어 전반에서 Lactococcus가 낮은 수준으로 검출되었고, 진저 비어에서 는 셀룰로오스를 분해하는 Trabulsiella 균주가 발견되어 프로바이오틱스 가능성을 시사하였다. 본 연구는 진저 비 어의 미생물 군집에 대한 최초의 연구로, 진저 비어 제조 시 재료로부터 기원한 미생물이 어떻게 변화하는지에 대 한 통찰을 제공한다. 또한, 다양한 환경에서의 발효 조건 이 미생물 군집과 제품의 품질에 미치는 영향을 탐구하는 데 기여할 것이다. 연구 결과는 진저 비어의 품질 향상에 대한 향후 연구에 중요한 자료를 제공할 것이다.

Refractory materials, such as molybdenum and niobium, are potential candidates for cladding material due to their high melting temperatures and desirable mechanical properties at higher temperatures than those of zirconium alloys. However, refractory materials have low resistance to oxidation at elevated temperatures. Therefore, this study examined silicide or aluminide surface coatings as protection against rapid oxidation of refractory materials at elevated temperatures for a potential accident-tolerant fuel cladding. Silicide or aluminide layers were formed on refractory metal substrates by using the pack cementation method. The steam oxidation behavior of both coated and uncoated samples was compared by thermogravimetric analysis at 1200°C. The weight changes of the coated samples were greatly reduced than those of uncoated samples. Microstructural analyses demonstrated that the silicide and aluminide layers were oxidized to form a protective surface oxide that prevented rapid oxidation of the refractory substrate at elevated temperatures.

High-entropy alloys (HEAs) represent a revolutionary class of materials characterized by their multi-principal element compositions and exceptional mechanical properties. Powder metallurgy, a versatile and cost-effective manufacturing process, offers significant advantages for the development of HEAs, including precise control over their composition, microstructure, and mechanical properties. This review explores innovative approaches integrating powder metallurgy techniques in the synthesis and optimization of HEAs. Key advances in powder production, sintering methods, and additive manufacturing are examined, highlighting their roles in improving the performance, advancement, and applicability of HEAs. The review also discusses the mechanical properties, potential industrial applications, and future trends in the field, providing a comprehensive overview of the current state and future prospects of HEA development using powder metallurgy.

This study investigated the growth dynamics of Lilium miquelianum bulb scales cultured in four different media formulations: Full Murashige-Skoog (FMS), Half Murashige-Skoog (HMS), Anderson Vitamin (AV), and Knudson Vitamin (KV). Bulb scales were collected from native populations on Jeju Island, Korea, and cultured in vitro for 12 weeks. Growth parameters, including bulb area and scale count, were measured at 3-week intervals. Statistical analyses revealed significant differences in growth rates across media types and time points. FMS consistently demonstrated the highest growth rates and final bulb area and scale count, while KV exhibited the lowest performance. Heatmap analysis showed that FMS achieved the highest weekly growth rates in most time points, with HMS showing comparable performance during early growth stages. Individual sample growth rates varied within media, highlighting the influence of both external and internal factors on growth. Correlation analysis revealed a positive association between bulb size and scale quantity, with FMS exhibiting the strongest relationship. These findings emphasize the importance of appropriate media selection for optimal L. miquelianum propagation, with FMS being the most suitable for extended cultivation. This research provides valuable insights for improving propagation efficiency and conservation efforts of L. miquelianum.

곤충호텔은 곤충이 서식할 수 있도록 인위적으로 만든 구조 물로서 정원이나 텃밭 등 다양한 곳에서 활용되고 있다. 하지만 식생유형, 재료, 설치 방식 등에 따른 곤충 유입 효과에 대한 검증은 부족하며, 국가 또는 지역에 따른 곤충의 종류와 생육환 경이 다름에도 불구하고 곤충호텔에 대한 국내 연구는 전무한 실정이다. 따라서 본 연구에서는 국립수목원을 대상으로 구체적 인 식생유형 및 재료에 따른 유입 곤충 특성을 확인하고, 국내 실정에 맞는 효과적인 곤충호텔 설치 및 관리방안에 대한 기준 을 제시하고자 하였다. 조사구는 세 가지 식생유형(초지, 침엽수 림, 활엽수림)으로 구분하였으며, 각 식생의 조사구에 두 가지 재료(참나무, 잣나무)를 활용한 곤충호텔을 설치하였다. 조사는 2년동안(2022~2023) 진행되었으며, 매년 4~9월까지 주 1회 씩(총 48회) 곤충호텔의 유입 곤충을 직접 채집하였다. 곤충호 텔에서 채집된 곤충은 총 9목 46과 129종 3,057개체로, 2022 년에는 7목 34과 85종 1,750개체, 2023년에는 8목 35과 77종 1,307개체가 출현하였다. 연도별로 비교하면 1차년도에 유입 된 곤충의 개체수가 2차년도 보다 약 1.3배 많았고, 재료의 부식 에 따른 곤충의 구성도 달라졌다. 식생유형에 따라 구분하면 기간에 상관없이 활엽수림에서 유입 곤충이 가장 많았고, 침엽 수림에서 가장 적었다. 또한 재료에 따른 구분에서는 참나무가 잣나무보다 유입 곤충이 많았다. 참나무의 경우 유입 곤충의 연도별 차이가 적은 반면, 잣나무는 1년차에 비해 2년차에 약 2.3배 줄었다. 상관분석과 계층적 군집분석을 통한 곤충의 유입 특성은 식생유형보다는 재료의 영향이 큰 것으로 확인되었고, 식생유형만 비교했을 경우 초지에 비해 활엽수림과 침엽수림의 유사성이 높았다. 결론적으로 생물 다양성 증진을 목적으로 곤 충호텔을 설치하고자 한다면 초지나 활엽수림에 참나무 재료를 사용하는 것이 유리하며, 잣나무 재료를 이용한다면 1년 주기로 재료 교체가 필요할 것으로 판단된다. 하지만 특정 재료를 선호 하는 곤충의 기주특이성이 확인되었기 때문에 다양한 종류의 재료를 같이 사용하는 것도 하나의 방법이라고 사료된다. 향후 이를 기반으로 전시원에서 실질적으로 활용할 수 있는 곤충호텔 모델 개발로 확대해 나갈 예정이다.

국립수목원 전시원의 지속가능하고 친환경적인 관리를 위하 여 유기물 멀칭재 처리에 대한 잡초 억제 효과를 조사하였다. 대표적 잡초인 쑥군락과 바랭이군락을 대상으로 고정방형구 (1×1㎡)를 3반복 설치하였다. 처리구의 유기물 멀칭재는 5가지 재료로 활엽수 우드칩, 침엽수 바크, 갈참나무 낙엽, 메타세쿼이 아 낙엽, 칠엽수 과피를 사용하였다. 멀칭재 중에서 갈참나무 낙엽, 활엽수 우드칩, 칠엽수 과피가 비교적 잡초 억제 효과가 높은 것으로 나타났다. 갈참나무 낙엽은 지중 온도를 낮게 유지 하는 효과가 상대적으로 좋았기 때문에 토양의 보습 효과가 뛰 어날 것으로 판단된다. 유기물 멀칭재 처리는 일년생식물의 발 생은 억제하고, 지중식물 및 반지중식물의 생육에 도움을 주는 효과가 있었다. 멀칭재의 무게보다 재료에 의한 빛 차단 면적이 잡초를 억제하는 주된 요인이었다. 본 연구 결과는 국립수목원 전시원에서 발생하는 부산물을 활용한 친환경적 관리 방안 수립 에 활용될 것이며, 향후 잡초의 발생 시기에 따른 멀칭 시기 및 유기물 멀칭재의 유지 관리에 대한 추가적인 연구가 필요하다.

Humic acids (HA), with their irregular polymeric structures and largely existing in grassland, present challenges in quality control due to significant variations in biological activities depending on extraction sources. To address this, we explored industrial byproducts as potential alternatives mimicking HA-like bioactivities. This study evaluates sulfite lignin, a byproduct of the pulp industry, as an eco-friendly biostimulant for enhancing plant growth and stress tolerance. Sulfite lignin demonstrated HA-like bioactivities, promoting seed germination and salt stress tolerance in Arabidopsis thaliana. Germination assays revealed that sulfite lignin significantly improved radicle and cotyledon emergence, particularly at low concentrations (8.6 mg L⁻¹), outperforming HA and kraft lignin. Additionally, under salt stress conditions, sulfite lignin-treated plants exhibited healthier phenotypes and maintained higher chlorophyll content compared to control treatments, similar to HA and kraft lignin. The findings highlight sulfite lignin as a promising, sustainable, and cost-effective biofertilizer, effectively replicating HA's biological functions while leveraging industrial byproducts.

Afoxolaner is an insecticide and acaricide that belongs to the isoxazoline chemical compound group. it has been used as an active pharmaceutical ingrdient in veterinary medicine to treat fleas and ticks in dogs. When patents expire between 2026 and 2066, it is expected that many products will be applied for approval as generic products, and reserch to establish accurate quality control methods must be conducted and managed. HPLC method was developed for the quantitative and qualitative of afoxolaner in veterinary medicinal products. The separation of active constituents for afoxolaner was achieved on a RP18 (4.6 x 150 mm, 5 μm) column using Water : Acetonitrile : MeOH (25:30:45 v/v/v) as mobile phase, with UV detection at 245 nm. The method was validated for specificity, linearity, accuracy and precision. All calibration curves showed good linearity (R2 of 0.999) within the concentration ranges (12.5 to 400 μg/mL). For accuracy, the recovery rate was calculated by spiking three concentrations of standard into the sample blank. The recovery rate was calculated to be 99.70~100.58%. Precision was measured 9 times repeatedly through intra-day, inter-day tests using standard. It showed excellent precision by satisfying the relative standard deviation of less than 2% both intra-day and inter-day. Limit of detection (LOD) and limit of quantitation (LOQ) were 2.0 μg/mL and 6.1 μg/mL, respectively. This method was successfully applied to analyzing afoxolaner drugs distributed in Korea. The HPLC method described in this study is accurate and reproducible and could be applied for the analysis of veterinary drugs of afoxolaner.

Silage inoculants, which include beneficial microorganisms like lactic acid bacteria (LAB), play a vital role in modern silage production by enhancing fermentation quality. This study evaluated the effectiveness of various commercial inoculants on the fermentation dynamics of Italian ryegrass silage over 45 days. The treatments included a control group and five inoculant formulations: T1 (Lactiplantibacillus plantarum), T2 (Lactiplantibacillus plantarum and Pediococcus pentosaceus), T3 (Lactiplantibacillus plantarum and Pediococcus pentosaceus and Lactiplantibacillus buchneri), T4 (Lactiplantibacillus plantarum and Lactiplantibacillus acidophilus and Lactiplantibacillus bulgaricus), and T5 (Lactiplantibacillus plantarum and Pediococcus pentosaceus and Enterococcus faecium). After 45 days, all treatment groups exhibited significantly higher crude protein (CP) content compared to the control group (80.64 g/kg dry matter (DM), p<0.05). Treatments T2 and T5, which incorporated combinations of Lactiplantibacillus plantarum, Pediococcus pentosaceus and Enterococcus faecium, showed higher CP contents at 105.53 and 107.05 g/kg DM, respectively. The inoculated silages also demonstrated a rapid pH reduction within the early days, with Lactiplantibacillus plantarum in T1 reducing the pH to 4.0 within four days. Additionally, inoculated treatments had significantly higher lactic acid levels than the control (67.96 g/kg DM, p<0.05), and T3 (Lactiplantibacillus buchneri) produced higher acetic acid levels (16.07 g/kg DM, p<0.05) than other inoculants. The control group also had a notably higher ammonia nitrogen content. In conclusion, while single-strain inoculants like Lactiplantibacillus plantarum are effective for rapid acidification, the use of combined bacterial strains can further enhance silage quality by improving lactic acid fermentation and nutrient preservation, particularly in treatments like Lactiplantibacillus plantarum and Pediococcus pentosaceus and Lactiplantibacillus buchneri and Enterococcus faecium.

Background: South Korea has recently faced record-high temperatures, which have adversely affected dairy production. Holstein cows, the primary dairy breed globally, are particularly sensitive to heat stress. In contrast, Jersey cows have shown greater heat tolerance, as demonstrated by phenotypic studies. Methods: We investigated physiological and molecular responses to heat stress in Holstein and Jersey cows by measuring rectal temperature, milk yield, and average daily gain, confirming Holstein cows’ greater vulnerability. To explore molecular mechanisms, we analyzed circulating microRNA profiles from whole blood samples collected under heat stress and normal conditions using microRNA-sequencing. Differential expression patterns were compared between the two breeds to identify biological pathways associated with heat stress. Results: Four microRNAs (bta-miR-20b, bta-miR-1246, bta-miR-2284x, and bta-miR- 2284y) were significantly differentially expressed in both breeds under heat stress (|FC| ≥ 2, p < 0.05). Notably, bta-miR-20b and bta-miR-1246 were linked to corpus luteum function and progesterone biosynthesis, while bta-miR-2284x and bta-miR- 2284y were associated with immune responses. A comparison of 11 potential heat stress-related microRNAs identified in previous studies of Holstein cows revealed consistent expression trends in Jersey cows, albeit with lower fold changes, suggesting their superior heat resilience. Conclusions: Our study highlights the physiological and microRNA-based differences in heat stress responses between Holstein and Jersey cows. Jersey cows exhibited greater resilience, supported by more stable microRNA expression profiles and improved heat stress indicators, making them a promising breed for dairy production in increasingly hot climates.

Following the previous study, which investigated the pharmacological properties of the Technekitty injection (Tc-99m), the toxicity of a single intravenous administration of the Technekitty injection (Tc-99m) and the side effects that may occur at the diagnostic dose were confirmed. The Technekitty injection (Tc-99m) was administered intravenously once at a dose of 0, 0.67, 2.0, and 6.0 mCi/kg to 5 male and female rats per group. Mortality, general symptom observation, and weight measurement were performed for 2 weeks, followed by observation of autopsy findings. There were no deaths, and no statistically significant weight change was observed. No abnormal systemic signs related to the Technekitty injection (Tc-99m) were observed. These results confirmed that Technekitty injection (Tc-99m) can be safely administered intravenously at doses up to 6.0 mCi/kg. Additionally, technetium-99m at an average dose of 2 mCi (74 MBq) has been verified as a diagnostic dose without adverse effects, allowing the Technekitty injection (Tc-99m) to be used safely without side effects at this dosage. This study demonstrates that the Technekitty injection (Tc-99m) has a wide safety margin, supporting its potential for clinical application. Moreover, these findings align with the nonclinical safety standards for radiopharmaceuticals, reinforcing its utility in veterinary medicine. The Technekitty injection (Tc-99m) is expected to be applicable for clinical diagnosis as a veterinary drug in Korea.

Background: The widespread use of smartphones and personal computers has contributed to a rise in thoracic kyphosis, a condition characterized by excessive outward curvature of the upper back. This condition can lead to reduced lung function, poor posture, and decreased spinal angles, all of which can cause reductions in forced vital capacity (FVC) and forced expiratory volume in 1 second (FEV1). Furthermore, these issues are often associated with scapular protraction and anterior tilting. Objects: This study aimed to investigate the effects of using a scapular support pillow in people with thoracic kyphosis, as well as determine the changes in respiratory capacity and postural alignment. Methods: Forty-one participants (25 males and 16 females) with thoracic kyphosis (Cobb angle > 40°) were included in the study. Their respiratory function (FVC, FEV1) and body posture (spinal angle, chest expansion, acromion-to-the-wall index) were measured before and after using the scapular support pillow for approximately 30 minutes. FVC, FEV1, spinal angle, chest expansion, and acromion-to-the-wall index were analyzed using paired-t test. The level of statistical significance was set at p < 0.05. Results: The results demonstrated significant improvements in all measured parameters. Both respiratory function and posture-related metrics showed notable increases after using the scapular support pillow. Conclusion: The use of a scapular support pillow can effectively improve respiratory function and postural alignment in patients with thoracic kyphosis. Our research makes a meaningful contribution by proposing an effortless and convenient treatment option for individuals with thoracic kyphosis.

Thyroid scanning using technetium-99m (99mTc) is the gold standard for diagnosing feline hyperthyroidism. In cats with an overactive thyroid, a thyroid scan is the most appropriate imaging technique to detect and localize any hyperfunctional adenomatous thyroid tissue. In this study, the pharmacological properties of the Technekitty injection (Tc-99m), developed as a diagnostic agent for feline hyperthyroidism using 99mTc as an active ingredient, were tested in FRTL-5 thyroid follicular cell line and ICR mice. The percentage of cell uptake of the Tc-99m in FRTL-5 thyroid cells was 0.182 ± 0.018%, which was about 6 times higher compared to Clone 9 hepatocytes. This uptake decreased by 38.2% due to competitive inhibition by iodine (sodium iodide). In tissue distribution tests by using ICR mice, the highest distribution was observed in the liver, kidneys, spleen, lungs, and femur at 0.083 hours after administration, and this distribution decreased as the compound was excreted through the kidneys, the primary excretory organ. Maximum distribution was confirmed at 1 hour in the small intestine, 6 hours in the large intestine, and 2 hours in the thyroid gland. Additionally, the total amount excreted through urine and feces over 48 hours (2 days) was 78.80% of the injected dose, with 37.70% (47.84% of the total excretion) excreted through urine and 41.10% (52.16% of the total excretion) through feces. In conclusion, the Tc-99m has the same mechanism of action, potency, absorption, distribution, metabolism, and excretion characteristics as 99mTc used for feline hyperthyroidism in the United States, Europe, and other countries, because the Technekitty injection (Tc-99m) contains 99mTc as its sole active ingredient. Based on these results, the Technekitty injection (Tc-99m) is expected to be safely used in the clinical diagnosis of feline hyperthyroidism.

Micro display technology has become a key element of digital innovation, gaining attention with immersive technologies like AR/VR, and advancing in fields such as education, military, and healthcare. This study aims to analyze micro display technology trends, which are increasingly important across industries, and to provide insights that help companies develop strategies for related technologies. Using the patent database, BERTopic, a recent topic modeling technique, was applied to identify key themes in micro display technology development. Time-series analysis was also conducted to examine changes in research trends over time. These findings are expected to guide companies and research institutions in setting future technology development directions and strategies.

On the wooden slabs unearthed from the tomb of Guojianao (formerly, Guo) in the period of Sun Wu of the Three Kingdoms, there is a sentence with the words “針囊一枚, 白糸一 , 青糸一”. Past scholars interpreted the three characters “ ” ,“ ” and “ ” as the word “縱 (zong)” , which means “silk thread”, and “ ” ,“ ” as the word “㢧 (卷, juan, volume)”, which is a measure word. This study verifies that the word “ ” is “ ”, and “ ”, “ ” are “ ”, which are all variants of the word “線 (xian, line)”. This was achieved by changing the phonetic note. The measure word “ ”, “ ” used for weighing silk thread should be a variant form of the word “斤 (jin)”, which is similar to the writing of “ (jin)” or “ (jin)” in the Pan’s clothing list of the Eastern Jin Dynasty.

As the Fourth Industrial Revolution advances, smart factories have become a new manufacturing paradigm, integrating technologies such as Information and Communication Technology (ICT), the Internet of Things (IoT), Artificial Intelligence (AI), and big data analytics to overcome traditional manufacturing limitations and enhance global competitiveness. This study offers a comprehensive approach by evaluating both technological and economic performance of smart factory Research and Development (R&D) projects, addressing gaps in previous studies that focused narrowly on either aspect. The research combines Latent Dirichlet Allocation (LDA) topic modeling and Data Envelopment Analysis (DEA) to quantitatively compare the efficiency of various topics. This integrated approach not only identifies key research themes but also evaluates how effectively resources are utilized within each theme, supporting strategic decision-making for optimal resource allocation. Additionally, non-parametric statistical tests are applied to detect performance differences between topics, providing insights into areas of comparative advantage. Unlike traditional DEA methods, which face limitations in generalizing results, this study offers a more nuanced analysis by benchmarking efficiency across thematic areas. The findings highlight the superior performance of projects incorporating AI, IoT, and big data, as well as those led by the Ministry of Trade, Industry, and Energy (MOTIE) and small and medium-sized enterprises (SMEs). The regional analysis reveals significant contributions from non-metropolitan areas, emphasizing the need for balanced development. This research provides policymakers and industry leaders with strategic insights, guiding the efficient allocation of R&D resources and fostering the development of smart factories aligned with global trends and national goals.