http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
오늘 본 자료
최환석(Hoan-Suk Choi),팽전(Qian Peng),지윤정(Yun-Jeong Ji),고건호(Keun-Ho Ko),이우섭(Woo-Seop Rhee) 한국디지털콘텐츠학회 2019 한국디지털콘텐츠학회논문지 Vol.20 No.1
It is very dangerous to lose consciousness or to prevent normal movement due to injury after a fall. In the dangerous situation, immediate response from surrounding people, protector and medical personnel is required. In order to recognize this complex situation, a data-fusion technique that utilizes various sensor data and related information together is required. In this paper, we propose data-fusion based dangerous-situation aware mechanism that combine users various movement data with previous context data. The proposed mechanism collects data every 100 ms to classify states as activity, falls, driving, and dangerous. Also it classifies activity state into five sub-state. We analyzed the performance of the proposed mechanism through experimentation. The proposed mechanism will be utilized in dangerous situation alert services.
최환석(Hoan-Suk Choi),팽전(Qian Peng),이우섭(Woo-Seop Rhee) 한국디지털콘텐츠학회 2020 한국디지털콘텐츠학회논문지 Vol.21 No.2
Recently, due to the changes in dietary life style and developments of the Internet technology, the app-based food ordering services have been emerged, and the delivery market continues to grow. Also, recommendation system is being introduced in a variety of areas because users want to easily find preferred targets. A recommendation system requires sufficient data including various attributes to provide stable and reliable performance. Thus, this paper proposes the data feature extension mechanism and machine learning based restaurant recommendation system to provide recommendations various purposes and situations. The proposed system provides the natural language comment based restaurant feature extraction method, the K-means based review score similar group generation method and the linear regression based ordering quantity prediction method to generate recommendation lists. Also, this paper implemented the proposed system by Python and conducted experiment of generating recommendation list based on restaurant review data of Guangzhou. The proposed system can flexibly reflect the feature of the target and allow for more detailed user taste. Also, it can predict the order trend because the order quantity prediction shows a margin of error is less than 20%.