• 한국임상수의학회지
• /
• 제40권4호
• /
• pp.243-259
• /
• 2023

Machine learning and deep learning (ML/DL) algorithms have been successfully applied in medical practice. However, their application in veterinary medicine is relatively limited, possibly due to a lack in the quantity and quality of relevant research. Because the potential demands for ML/DL applications in veterinary clinics are significant, it is important to note the current gaps in the literature and explore the possible directions for advancement in this field. Thus, a scoping review was conducted as a situation analysis. We developed a search strategy following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. PubMed and Embase databases were used in the initial search. The identified items were screened based on predefined inclusion and exclusion criteria. Information regarding model development, quality of validation, and model performance was extracted from the included studies. The current review found 55 studies that passed the criteria. In terms of target animals, the number of studies on industrial animals was similar to that on companion animals. Quantitative scarcity of prediction studies (n = 11, including duplications) was revealed in both industrial and non-industrial animal studies compared to diagnostic studies (n = 45, including duplications). Qualitative limitations were also identified, especially regarding validation methodologies. Considering these gaps in the literature, future studies examining the prediction and validation processes, which employ a prospective and multi-center approach, are highly recommended. Veterinary practitioners should acknowledge the current limitations in this field and adopt a receptive and critical attitude towards these new technologies to avoid their abuse.

• International Journal of Automotive Technology
• /
• 제5권2호
• /
• pp.95-108
• /
• 2004

Lanekeeping assistance has the potential to save thousands of lives every year by preventing accidental road departure. This paper presents experimental validation of a potential field lanekeeping assistance system with quantitative performance guarantees. The lanekeeping system is implemented on a 1997 Corvette modified for steer-by-wire capability. With no mechanical connection between the hand wheel and road wheels the lanekeeping system can add steering inputs independently from the driver. Implementation of the lanekeeping system uses a novel combination of a multi-antenna Global Positioning System (GPS) and precision road maps. Preliminary experimental data shows that this control scheme performs extremely well for driver assistance and closely matches simulation results, verifying previous theoretical guarantees for safety. These results also motivate future work which will focus on interaction with the driver.

• Journal of electromagnetic engineering and science
• /
• 제13권3호
• /
• pp.192-194
• /
• 2013

This paper proposes a validation model using a reverberation chamber for verification of the effectiveness and accuracy of the statistical power balance (PWB) method. The PWB method treats electromagnetic (EM) field distributions in certain environments as a representative value of an "ensemble" average, so a reverberation chamber is utilized as a testing apparatus to experimentally simulate the sets of measurement field "ensemble" inside a complex environment. Measurements are performed with various test conditions for environmental loss using flat absorbers, and front and side doors of the chamber. Comparison between the PWB analysis and the measurement results shows good agreement, indicating the usefulness of this statistical analysis method for stating the electromagnetic field inside the EM zone as a representative value to general structures.

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 2004년도 Proceedings of ISRS 2004
• /
• pp.302-305
• /
• 2004

The aim of this study is to apply and crossvalidate a spatial logistic multiple-regression model at Boun, Korea, using a Geographic Information System (GIS). Landslide locations in the Boun area were identified by interpretation of aerial photographs and field surveys. Maps of the topography, soil type, forest cover, geology, and land-use were constructed from a spatial database. The factors that influence landslide occurrence, such as slope, aspect, and curvature of topography, were calculated from the topographic database. Texture, material, drainage, and effective soil thickness were extracted from the soil database, and type, diameter, and density of forest were extracted from the forest database. Lithology was extracted from the geological database and land-use was classified from the Landsat TM image satellite image. Landslide susceptibility was analyzed using landslide-occurrence factors by logistic multiple-regression methods. For validation and cross-validation, the result of the analysis was applied both to the study area, Boun, and another area, Youngin, Korea. The validation and cross-validation results showed satisfactory agreement between the susceptibility map and the existing data with respect to landslide locations. The GIS was used to analyze the vast amount of data efficiently, and statistical programs were used to maintain specificity and accuracy.

• 한국경영정보학회:학술대회논문집
• /
• 한국경영정보학회 2007년도 International Conference
• /
• pp.557-561
• /
• 2007

Along with the significance of information systems in today s global business operation, the significance of information systems control and audit is ever increasing in the effort to secure accuracy and integrity of vital business data. A study is undertaken to integrate Food and Drug Administration computerized systems validation regulations. Securities & Exchange Commission Sarbanes-Oxley Act of 200 and other significant regulations, and lastly, People Capability Maturity Model into one comprehensive information system validation model. The initial benefits to this comprehensive model are convenience, time-saving, and synchronization of the regulations. An organization that is striving for a high level of quality system in its essential operating areas of organization may opt for this model. After the complete development of the model, a field test would be scheduled to test its efficacy and validity.

• Journal of Community Nutrition
• /
• 제5권2호
• /
• pp.83-92
• /
• 2003

Several self-administered dietary assessment questionnaires have recently been developed, validated, and used in nutritional epidemiological and clinical studies in Japan. This article describes recent evidence on development and validation of them. After extensive search of published articles both in English and Japanese languages, we identified 25 articles on 13 questionnaires of which validation studies have existed. Number of foods/menus assessed varied from 31 to 169 according to questionnaires. Eleven questionnaires were food frequency type, either with fixed portion size or semiquantitative, and two diet history types. All the 13 questionnaires were validated against intakes assessed with dietary record or 24-hour recall, and only two with biomarkers. Number of subjects used in the studies was between 23 and 350. All the studies used adult subjects. In the studies with dietary record or recall, the correlation coefficient for or orgy intake was between 0.22 and 0.65 (median = 0.44). Median correlation coefficient for nutrients was between 0.21 and 0.61. In the studies with biomarkers, serum marine-origin n-3 polyunsaturated fatty acids and carotenes, and urinary potassium seemed useful biomarkers. In conclusion, recent progress of this field in Japan is remarkable. But more research is needed for validation studies with biomarkers, and the development and validation of questionnaires for children and elderly subjects. (J Community Nutrition 5(2) : 83∼92,2003)

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 2004년도 Proceedings of ISRS 2004
• /
• pp.298-301
• /
• 2004

The aim of this study is to cross-validate of spatial probability model, artificial neural network at Boun, Korea, using a Geographic Information System (GIS). Landslide locations were identified in the Boun, Janghung and Youngin areas from interpretation of aerial photographs, field surveys, and maps of the topography, soil type, forest cover and land use were constructed to spatial data-sets. The factors that influence landslide occurrence, such as slope, aspect and curvature of topography, were calculated from the topographic database. Topographic type, texture, material, drainage and effective soil thickness were extracted from the soil database, and type, diameter, age and density of forest were extracted from the forest database. Lithology was extracted from the geological database, and land use was classified from the Landsat TM image satellite image. Landslide susceptibility was analyzed using the landslide­occurrence factors by artificial neural network model. For the validation and cross-validation, the result of the analysis was applied to each study areas. The validation and cross-validate results showed satisfactory agreement between the susceptibility map and the existing data on landslide locations.

• 한국대기환경학회지
• /
• 제14권3호
• /
• pp.219-228
• /
• 1998

In the present investigation, a numerical model developed for the prediction of the wind flow over complex terrain is validated by comparing with the field experiments. For the solution of the Reynolds - Averaged Clavier- stokes equations which are the governing equations of the microscale atmospheric flow, the model is constructed based on the finite-volume formulation and the SIMPLEC pressure-correction algorithm for the hydrodynamic computation. The boundary- fitted coordinate system is employed for the detailed depiction of topography. The boundary conditions and the modified turbulence constants suitable for an atmospheric boundary- layer are applied together with the k- s turbulence model. The full- scale experiments of Cooper's Ridge, Kettles Hill and Askervein Hill are chosen as the validation cases . Comparisons of the mean flow field between the field measurements and the predicted results show good agreement. In the simulation of the wind flow over Askervein Hill , the numerical model predicts the three dimensional flow separation in the downslope of the hill including the blockage effect due to neighboring hills . Such a flow behavior has not been simulated by the theoretical predictions. Therefore, the present model may offer the most accurate prediction of flow behavior in the leeside of the hill among the existing theoretical and numerical predictions.

• 천문학논총
• /
• 제32권1호
• /
• pp.37-39
• /
• 2017

We present the first results of a new data analysis pipeline for processing extragalactic AKARI/IRC images. The main improvements of the pipeline over the standard analysis are the removal of Earth shine and image distortion correction. We present the differential number counts of the AKARI/IRC S11 filter in the IRAC validation field. The differential number counts are consistent with S11 AKARI NEP deep and $12{\mu}m$ WISE NEP number counts, and with a phenomenological backward evolution galaxy model, at brighter fluxes densities. There is a detection of fainter galaxies in the IRAC validation field.

• 한국항공우주학회지
• /
• 제37권2호
• /
• pp.209-214
• /
• 2009

본 논문에서는 풍력발전시스템의 정상 및 과도 응답과 같은 동적 거동 분석은 물론 피치제어 시스템의 성능 예측을 위하여 개발된 시뮬레이션 소프트웨어의 검증을 위하여 실제 현장에서 측정한 자료와 비교 검증을 수행하였고 실제와 유사하도록 시스템 모델링 및 소프트웨어의 개선 방향을 제시함으로써 향후 본 연구를 통하여 개발한 소프트웨어의 유용성을 제시하고자 한다.