• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.16-18
• /
• 2023

본 연구는 궤적 데이터(trajectory data)를 대상으로 증강 그래프 기반의 그래프 뉴럴 네트워크를 활용하여 다음에 방문한 장소를 추천하는 모델을 제안한다. 제안 모델은 전체 궤적 데이터를 그래프로 표현하여 추출한 글로벌 궤적 플로우의 특성을 다음 방문할 POI 추천에 활용한다. 이때, POI 추천시 자주 발생하는 두 가지 문제를 추가로 해결함으로써 POI 추천의 정확도를 높이는 것을 목표로 한다. 첫 번째 문제는 추천 대상 궤적 데이터의 길이가 짧은 경우에 성능 저하가 발생한다는 것이다. 두 번째 문제는 콜드-스타트 문제이다. 기존 POI 추천 모델은 매우 적은 방문 기록만 가지는 사용자 또는 POI에 대해서는 매우 낮은 예측 성능을 보인다. 본 연구에서는 궤적 그래프에서 일부 엣지를 삭제하여 생성한 증강 그래프 기반의 궤적 플로우 특징 기반 모델을 제안함으로써 짧은 길이의 궤적 데이터 및 콜드-스타트 사용자/POI에 대한 추천 성능을 높인다.

• International Journal of Highway Engineering
• /
• v.10 no.3
• /
• pp.129-138
• /
• 2008

The finite element method(FEM) was one of tools used to solve problem of previous Concrete Pavement and was applied to Korea Pavement Research Program Study. This study used the ABAQUS and the fortran analysis program to calculate the critical stress on jointed concrete pavement and compared and analyzed the results by using neural networks and linear regression model. In that case, which are not enough analysises by using FEM programs though many input variables, when the results of FEM with NN and linear regression models are compared, there are some differences. The other cases, which are reduced input variables and a lot of analysises each of them, results of Neural Networks(NN) and linear regression models are simulated to them of FEM. But, the result of NN is more exact than them of linear regression at the (0,0), (1,1). On the results of this study, it is suggested that the calculation of stress using NN is more compatible to Korea Pavement Research Program Study.

• Journal of the Korea Convergence Society
• /
• v.8 no.12
• /
• pp.49-53
• /
• 2017

Breast cancer is the sccond most female cancer patient in the entire female cancer patient, and has emerged as the highest contributor to female cancer deaths. If breast cancer id detected early, the cure rate is 92 percent. However, if early detection fails, breast cancer has a very high rate of metastasis. The transition from cancer to cancer has become more successful as cancer progresses. Early diagnosis of cancer is an important factor in improving quality of life. Examples of breast cancer include Mammograph, ultrasound, and Momotome. Mommography is not only painful for the examiner, but also for easy access to breast cancer exam inations. In this paper, breast cancer diagnosis data mammograph data was used. In addition, the Neural Network were classified for early diagnosis of breast cancer early using NEWFM. After learning of data using NEWFM, the accuracy of the breast cancer data classification was 84.4391%.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.17 no.4
• /
• pp.138-149
• /
• 2018

In this paper, we propose a system model which effectively mitigates impulsive noise that degrades the performance of power line communication. Recently, deep learning have shown effective performance improvement in various fields. In order to mitigate effective impulsive noise, we applied a convolution neural network which is one of deep learning algorithm to conventional system. Also, we used a successive interference cancellation scheme to mitigate impulsive noise generated from multi-users. We simulate the proposed model which can be applied to the power line communication in the Section V. The performance of the proposed system model is verified through bit error probability versus SNR graph. In addition, we compare ZF and MMSE successive interference cancellation scheme, successive interference cancellation with optimal ordering, and successive interference cancellation without optimal ordering. Then we confirm which schemes have better performance.

• Journal of Intelligence and Information Systems
• /
• v.29 no.2
• /
• pp.323-338
• /
• 2023

TV advertising with deep analysis of watching pattern of audiences is important to set-top box audience targeting. Applying session-based recommendation model(SBR) to internet commercial, or recommendation based on searching history of user showed its effectiveness in previous studies, but applying SBR to the TV advertising was difficult in South Korea due to data unavailabilities. Also, traditional SBR has limitations for dealing with user preferences, especially in data with user identification information. To tackle with these problems, we first obtain set-top box data from three major broadcasting companies in South Korea(SKB, KT, LGU+) through collaboration with Korea Broadcast Advertising Corporation(KOBACO), and this data contains of watching sequence of 4,847 anonymized users for 6 month respectively. Second, we develop personalized session-based recommendation model to deal with hierarchical data of user-session-item. Experiments conducted on set-top box audience dataset and two other public dataset for validation. In result, our proposed model outperformed baseline model in some criteria.