• The Korean Journal of Applied Statistics
• /
• v.26 no.5
• /
• pp.835-845
• /
• 2013

Travel-time is considered the most typical and preferred traffic information for intelligent transportation systems(ITS). This paper proposes a real-time travel-time prediction method for a national highway. In this paper, the K-nearest neighbor(KNN) method is used for travel time prediction. The KNN method (a nonparametric method) is appropriate for a real-time traffic management system because the method needs no additional assumptions or parameter calibration. The performances of various models are compared based on mean absolute percentage error(MAPE) and coefficient of variation(CV). In real application, the analysis of real traffic data collected from Korean national highways indicates that the proposed model outperforms other prediction models such as the historical average model and the Kalman filter model. It is expected to improve travel-time reliability by flexibly using travel-time from the proposed model with travel-time from the interval detectors.

• Journal of the Korean Data and Information Science Society
• /
• v.20 no.5
• /
• pp.809-818
• /
• 2009

Collaborative filtering approach predicts the preference of active user about specific items transacted on the e-commerce by using others' preference information. To improve the prediction accuracy through collaborative filtering approach, it must be needed to gain enough preference information of users' for predicting preference. But, a bit much information of users' preference might wrongly affect on prediction accuracy, and also too small information of users' preference might make bad effect on the prediction accuracy. This research suggests the method, which decides suitable numbers of neighbor users for applying collaborative filtering algorithm, improved by existing k nearest neighbors selection methods. The result of this research provides useful methods for improving the prediction accuracy and also refines exploratory data analysis approach for deciding appropriate numbers of nearest neighbors.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.35 no.5
• /
• pp.259-266
• /
• 2022

In this study, we developed a deep learning and vision sensor-based assembly performance evaluation method isfor prefabricated steel structures. The assembly parts were segmented using a modified version of the receptive field block convolution module inspired by the eccentric function of the human visual system. The quality of the assembly was evaluated by detecting the bolt holes in the segmented assembly part and calculating the bolt hole positions. To validate the performance of the evaluation, models of standard and defective assembly parts were produced using a 3D printer. The assembly part segmentation network was trained based on the 3D model images captured from a vision sensor. The sbolt hole positions in the segmented assembly image were calculated using image processing techniques, and the assembly performance evaluation using the k-nearest neighbor algorithm was verified. The experimental results show that the assembly parts were segmented with high precision, and the assembly performance based on the positions of the bolt holes in the detected assembly part was evaluated with a classification error of less than 5%.

• Journal of Korean Society of Transportation
• /
• v.34 no.2
• /
• pp.158-167
• /
• 2016

This study evaluates potential sources of errors in k-NN(k-nearest neighbor) algorithm such as procedures, variables, and input data. Previous research has been thoroughly reviewed for understanding fundamentals of k-NN algorithm that has been widely used for short-term traffic states prediction. The framework of this algorithm commonly includes historical data smoothing, pattern database, similarity measure, k-value, and prediction horizon. The outcomes of this study suggests that: i) historical data smoothing is recommended to reduce random noise of measured traffic data; ii) the historical database should contain traffic state information on both normal and event conditions; and iii) trial and error method can improve the prediction accuracy by better searching for the optimum input time series and k-value. The study results also demonstrates that predicted error increases with the duration of prediction horizon and rapidly changing traffic states.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.23 no.1
• /
• pp.7-11
• /
• 2013

Taking experts' knowledge to recommend items has shown some promising results in recommender system research. In order to improve the performance of the existing recommendation algorithms, previous researches on expert-based recommender systems have exploited the knowledge of a common expert group for all users. In this paper, we study a problem of identifying personalized experts within a user group, assuming each user needs different kinds and levels of expert help. To demonstrate this idea, we present a framework for using Support Vector Machine (SVM) to find varying expert groups for users; it is shown in an experiment that the proposed SVM approach can identify personalized experts, and that the person-alized expert-based collaborative filtering (CF) can yield better results than k-Nearest Neighbor (kNN) algorithm.

• Science of Emotion and Sensibility
• /
• v.10 no.3
• /
• pp.289-297
• /
• 2007

Color image scale captures the knowledge of colorists and represents both adjectives and colors in the same adjective image scales in order to select color(s) corresponding to an adjective. Due to the difficulty of psychological experiment and statistical analysis, in general, only a limited number of colors are located in the color image scales. This can make color selection process hard especially to non-expert. In this paper, we propose an interpolation of color image scale based on the fuzzy K-nearest neighbor method, which provides continuous colors according to the coordinates of the image scales. The experimental results show that the interpolated image scales can be practically useful for color selection process.

• Journal of the Korea Society of Computer and Information
• /
• v.24 no.10
• /
• pp.79-89
• /
• 2019

While most research focuses on the k-nearest neighbors (kNN) queries in the database community, an important type of proximity queries called k-farthest neighbors (kFN) queries has not received much attention. This paper addresses the problem of finding the k-farthest neighbors in road networks. Given a positive integer k, a query object q, and a set of data points P, a kFN query returns k data objects farthest from the query object q. Little attention has been paid to processing kFN queries in road networks. The challenge of processing kFN queries in road networks is reducing the number of network distance computations, which is the most prominent difference between a road network and a Euclidean space. In this study, we propose an efficient algorithm called FANS for k-FArthest Neighbor Search in road networks. We present a shared computation strategy to avoid redundant computation of the distances between a query object and data objects. We also present effective pruning techniques based on the maximum distance from a query object to data segments. Finally, we demonstrate the efficiency and scalability of our proposed solution with extensive experiments using real-world roadmaps.

• Journal of KIISE:Databases
• /
• v.32 no.4
• /
• pp.429-440
• /
• 2005

As rapid improvement in electronic technologies makes sensor hardware more powerful and capable, the application range of sensor networks Is getting to be broader. The main purpose of sensor networks is to monitor the phenomena in interesting regions (e.g., factory warehouses, disaster areas, wild fields, etc) and return required data. The k Nearest Neighbor (KNN) query that finds k objects which are geographically close to the given point is an Important application in sensor networks. However, most previous approaches are either seem to be impractical or are not energy-efficient in resource-limited sensor networks. In this paper. we propose an efficient KNN query processing method in sensor networks. In the proposed method, we dynamically increase searching boundary, if necessary, and traverse nodes inside the boundary until finding k nearest neighbors. Since only the representative sensor nodes are visited, our algorithm reduces a number of messages. We show thorough experiments that the proposed method performs better than the existing method in various network environments.

• Proceedings of the Korean Information Science Society Conference
• /
• 2012.06c
• /
• pp.125-127
• /
• 2012

현재 이종 그래프에 대한 연구가 활발히 진행되고 있다. 특히 추천 및 검색 분야에서 이종 그래프를 활용하여 성능을 높이는 성과가 두드러진다. 이종 그래프는 다양한 정보를 갖고 있으며, 특히 2단계 랜덤 워크 확률은 여러 유용한 정보를 가지고 있다. "어떤 사용자가 많이 본 영화를 많이 본 사용자", "어떤 사용자의 이웃이 많이 구입한 상품" 등이 그예이다. 하지만 이러한 정보를 실시간에 계산하기는 어려우며, 미리 계산해두는 것도 시간이 많이 든다. 이에 따라, 본 연구에서는 모든 출발 노드-도착 노드 쌍에 대한 2단계 랜덤 워크를 빠르게 미리 계산하는 알고리듬을 제시한다. 동일한 이웃 노드를 다수 가진 두 노드에서 출발하는 랜덤 워크 확률 값은 서로 비슷하다는 사실을 이용하여, 이전 계산 결과를 활용하여 근접 노드 목록에 대한 임의 접근 횟수를 줄인다. 더불어 본 알고리듬과 관련된 현안을 몇 가지 소개한다.

• KIPS Transactions on Software and Data Engineering
• /
• v.3 no.6
• /
• pp.215-218
• /
• 2014

In hashing-based methods for approximate nearest neighbors(ANN) search, by mapping data points to k-bit binary codes, nearest neighbors are searched in a binary embedding space. In this paper, we present a hashing method using a PCA-based clustering method, Principal Direction Divisive Partitioning(PDDP). PDDP is a clustering method which repeatedly partitions the cluster with the largest variance into two clusters by using the first principal direction. The proposed hashing method utilizes the first principal direction as a projective direction for binary coding. Experimental results demonstrate that the proposed method is competitive compared with other hashing methods.