http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Donggeun Kim,Juyong Ko,Minho Sun,Jai Woo Lee 한국경영정보학회 2023 한국경영정보학회 학술대회논문집 Vol.2023 No.11
These days, business intelligence has witnessed the various challenges of big data analytics due to exponentially growing information with uncertainty existing in the market. To effectively analyze the large amount of data from various sources in business, algorithms of artificial intelligence techniques should be efficiently improved. We present a comprehensive and novel approach for the evaluation of outcome by using the interactions of features and then applying this approach to estimate the trend by jointly modelling features in management data. In terms of predictive accuracy, the proposed method outperformed machine learning methods such as regression, penalized regression, decision tree, random forest and k-nearest neighbors in the high-dimensional business data analysis. Data-preprocessing was used to curate the data for better prediction and network analysis was conducted to appropriately visualize and analyze the data analysis results. The business literature represents that investigating artificial intelligence techniques with theoretical ideas for big data analytics can have an impact on reducing costs and risks in management. Future directions have been devised to elucidate the gap between actual values in real-world data of business intelligence and predicted values by the proposed approach. Machine learning methods including features of demographic and strategic data can estimate the effect of marketing characteristics. Using the proposed method, businesses may better assess strongly correlated features with the target output in the similarly structured business data.
Donggeun Kim,Sangwoo Park,Donggoo Kang,Joonki Paik 대한전자공학회 2020 IEIE Transactions on Smart Processing & Computing Vol.9 No.1
A single shot multibox detector (SSD) is used as a baseline for many object detection networks, since it can provide sufficiently high accuracy in real time. However, it cannot deal with objects of various sizes, because features used in an SSD are not robust to multi-scale objects. To solve this problem, we present an improved feature pyramid for using multi-scale context information. The proposed feature pyramid fuses only adjacent features of the conventional SSD to achieve high accuracy without decreasing the processing speed. Our detector, with a 320×320 input, achieved 79.1% mean average precision (mAP) at 63 frames per second on a Pascal Visual Object Classes Challenge 2007 test set using a single Nvidia 1080 Ti graphics processing unit. This result shows better performance than existing SSDs.
Robust Parameter Design of an Ascender Affecting Rope Deformation for High Repeatability
DongGeun Hyun,Sungjun Park,Jeongmo Yang,TaeWon Seo 한국정밀공학회 2023 International Journal of Precision Engineering and Vol.24 No.5
In this study, an optimized traction pulley is designed to achieve high repeatability of rope winches of surface-cleaning robots that estimate positions using the lengths of the ascender and flexible rope. When a slip occurs, measuring the length of the rope becomes difficult, which further complicates position estimation and control. However, if the slip is constant and repeatable, the error can be corrected by estimating the position. Herein, traction pulley’s design parameters(Pulley diameter, Number of groove,Pressure roller force) was experimentally evaluated using Taguchi method and Full Factorial Design. Based on two experiments, the optimal design parameters of the pulley were determined to be a diameter of 205 mm, 45 grooves, and a pressure roller force of 47 N. The optimized pulley demonstrated improved performance by reducing the standard deviation in the position error by 70.1%compared with that of the optimized pulley reported in a previous study. Thus, applying the optimized traction pulley in future studies is expected to improve the accuracy of position estimation.