인공지능(AI: Artificial Intelligence) 기법을 이용한 남자 테니스 선수 경기기록의 시각화

최형준(Hyong Jun Choi),권민혁(Min Hyuk Kwon) 한국체육측정평가학회 2012 한국체육측정평가학회지 Vol.14 No.3

본 연구는 남자 테니스 선수의 경기력을 나타내는 방법에 있어서, 스포츠데이터 분석에 적용되지 않았던 인공지능기법을 활용하여 남자 테니스 선수의 개인 경기기록을 시각화하고자 하였다. 본 연구에서는 남자 테니스 선수228명이 2005년부터 2010년까지 출전한 4대 그랜드슬램 대회 2,060세트에 대해서 최종 진출한 라운드(1∼7라운드)로 구분하여 Microsoft Excel 2010로 수집 및 정리하였으며, 자료의 시각화를 위하여 SOM toolbox가 적용된Matlab 14.0 소프트웨어를 이용하여 새몬의 매핑과 자기구성지도를 설계하였다. 새몬의 매핑은 3차원으로 설계하여 적용하였으며, 자기구성지도는 [17 13] 크기로 나타났으며, 훈련을 마친 후의 양자화 오차는 1.451, 지형화 오차는 0.066으로 나타났다. 본 연구를 통하여 얻은 결론은 첫째, 새몬의 매핑은 대용량 자료의 경우에는 시각화된자료를 해석하는데 어려움이 있었다. 둘째, 3차원 새몬의 매핑을 통하여 투영할 경우, x-y, x-z, y-z 좌표 간 비교를 통하여 자료를 재해석하는데 용이하였고, 셋째, 자기구성지도를 이용할 경우, 대용량 자료를 대표할 수 있는 뉴런을 이용하여 각 변인의 특성을 고려하여 자료처리 일련의 과정을 시각화 할 수 있었다. This study was to discuss between two mapping algorithm among artificial intelligence techniques using performances basing on official stats of men`s single tennis player. Totally, 2,060 sets data from 228 players from 2005 to 2010 tennis Grand slam competitions were selected and gathered into Microsoft Excel 2010. And a Matlab 14.0 version with Self-organizing map toolbox developed by Helsinki University of Technology was used to design the Sammon`s mapping and Self-organizing map projection. Sammon`s mapping techniques were designed as 3-dimensional structure and a size of the Self-organizing map was [17 13]. After the training of Self-organizing map, the quantization error was 1.451 and the topographic error was 0.066. Consequently, the results of the study was as following; Firstly, the Sammon`s mapping technique was unable to descript the details of data hardly when the size of data was huge. Secondly, 3-dimensional Sammon`s mapping technique was able to visualize data easily with x-y, x-z and y-z coordination graphs. Thirdly, using Self-organizing map brought an advantage of visualization for data process using comprehensive neurons which could apply variables` characteristics into the results.

A METHOD OF IMAGE DATA RETRIEVAL BASED ON SELF-ORGANIZING MAPS

Lee, Mal-Rey,Oh, Jong-Chul 한국전산응용수학회 2002 The Korean journal of computational & applied math Vol.9 No.2

Feature-based similarity retrieval become an important research issue in image database systems. The features of image data are useful to discrimination of images. In this paper, we propose the highspeed k-Nearest Neighbor search algorithm based on Self-Organizing Maps. Self-Organizing Maps (SOM) provides a mapping from high dimensional feature vectors onto a two-dimensional space. The mapping preserves the topology of the feature vectors. The map is called topological feature map. A topological feature map preserves the mutual relations (similarity) in feature spaces of input data. and clusters mutually similar feature vectors in a neighboring nodes. Each node of the topological feature map holds a node vector and similar images that is closest to each node vector. In topological feature map, there are empty nodes in which no image is classified. We experiment on the performance of our algorithm using color feature vectors extracted from images. Promising results have been obtained in experiments.

자기구성지도(Self-Organizing Map)를 이용한 엘리트 유소년 선수의 나이 예측 연구

정연성,최형준 한국스포츠학회 2019 한국스포츠학회지 Vol.17 No.4

The purpose of current study was to identify if physique and physical fitness data could predict the actual age of an elite youth athletes using the Self-Organizing Map among the artificial intelligent techniques into the field of sport science. In order to approach the purpose of this study, 4,432 elite athletes participated in the physical fitness test, and 1,889 elite youth athletes were finally selected. To derive the predict data, the proposed the study of Kohonen(1989) was used in Self-Organizing Map. Based on physique and physical fitness data of youth athletes, the results of the predicted accuracy of actual age and estimated age were: Predictive accuracy was around 87.3% for the same age±5, 85.0% for the same age±4, 81.2% for the same age±3, 75.2% for the same age±2, 51.8% for the same age±1, 25.5% for the same age. It is required that the following researches would be considered more data characteristics in order to using Self-Organizing Map to learn the characteristics of performance. 이 연구는 경기도 소재 엘리트 유소년 선수의 체격 및 체력자료를 바탕으로 인공지능 기법 중에 하나인 자기구성지도를 학습한 후, 체격 및 체력자료가 엘리트 유소년 선수의 실제 나이를 예측 할 수 있는지를 알아보는데 목적을 두었다. 이 연구의 목적을 달성하기 위하여 경기스포츠과학센터에서 실시한 체력측정에 임한 엘리트 선수 4,432명의 자료를 수집하였으며 수집된 전체 자료 중에서 결측값이 존재하는 사례를 제거하였으며, 최종적으로 선정된 자료의 수는 1,889명의 엘리트 유소년 선수의 체력 자료였다. 예측 자료의 도출을 위하여 Kohonen(1989)의 연구에서 제안된 인공지능 기법 중에 하나인 자기구성지도를 이용하였다. 유소년 선수들의 체격 및 체력 자료를 바탕으로 실제나이와 추정나이의 예측 정확성을 알아본 결과, 동일 연령±5는 약 87.3%, 동일 연령±4세는 약 85.0%, 동일 연령±3세는 약 81.2%, 동일 연령±2세는 약 75.2%, 동일 연령±1세는 약 51.8%, 동일 연령은 약 25.5%의 예측 정확성이 나타났다. 향후 연구를 통하여 경기력의 특성을 자기구성지도에 나타내게 학습시켜 보다 많은 자료의 특성이 고려될 수 있도록 해야 할 것으로 사료된다.

Application of Self-organizing Mapping-Random Forest Model to Map Landslide Susceptibility in Zigui Basin, Three Gorges Reservoir Region, China

( Changdong Li ),( Jingjing Long ),( Zhiyong Fu ),( Wenqiang Chen ) 대한지질공학회 2019 대한지질공학회 학술발표회논문집 Vol.2019 No.2

The Zigui basin is one of the most landslide-prone areas where thousands of landslides are distributed. Researches show that areas sharing the same conditions as identified landslides are clearly potential areas for future disasters. Performing the landslide susceptibility mapping is a heated issue in the area. Continued improvements in high-resolution satellite images, and the developments of unmanned aerial vehicles (UAVs) and site-investigation function well in constructing effective, high-quality landslide databases. GIS technology and machine learning algorithms have been widely applied in landslide susceptibility prediction. Whereas, whether the random and subjective selection of the landslides or non-landslides grid cells are reasonable in research of landslide susceptibility mapping is the existence problem. Based on the Two step cluster (TSC) algorithm and the Self-organizing mapping - Random forest (SOM-RF) model, a novel hybrid model is proposed to overcome the above drawbacks. SOM is used to produce a preliminary landslide susceptibility mapping. TSC algorithm is applied in telling apart the most reasonable True-Positive (TP) from recorded landslide grid cells in high-susceptibility zones and the False-Positive (FP) in low-susceptibility zones. Afterwards, the labeled datasets are imported into the RF model for training. And then the trained SOM-RF model is utilized to perform an improved landslide susceptibility mapping. Most areas with high or very high susceptibility are located within the hydro-fluctuation belt of the TGR. Compared with the susceptibility mapping produced by single RF model, the results of SOM-RF model demonstrate to have the superior prediction skill and higher reliability.

An Improved SOM Based Surface Texture Synthesis

Jin Zhao,Chen Deyun 보안공학연구지원센터 2015 International Journal of Signal Processing, Image Vol.8 No.12

This paper proposed a new method for surface texture synthesis using improved self-organizing maps as the synthesis logic. The method will first loop map the sample image to the target surface until the surface is full filled, then construct a improved self-organizing maps model, in which use sample image as the input level and the target three-dimensional surface as out put level, to adjust the pixel position in target surface, and the mapping area is controlled in the neighborhood of local extremum by reducing search area of variables. The new algorithm not only speed up the synthesis progress, and also enhances the image quality of target surface texture, and compare with the previous algorithm, it is not necessary for the user’s intervention. Compared with previous results for the proof of our concepts, we have successfully implemented the experimental results and the proposed algorithm.

한국 남자 국가대표 선수의 체격,체력검사 결과와 자기구성지도(Self-Organized Map)를 이용한 스포츠 종목의 군집분석 -사례연구를 중심으로

최형준 ( Hyong Jun Choi ),정연성 ( Yeon Sung Jung ),고병구 ( Byoung Goo Kob ) 체육과학연구원 2009 체육과학연구 Vol.20 No.3

이 연구는 자기구성지도와 체격·체력검사 결과를 토대로 한국 남자 국가대표 선수의 스포츠 종목을 군집해 보고, 군집된 스포츠 종목을 수행하는데 필요한 체력요인을 알아보는데 목적을 둔다. 자기구성지도는 경쟁학습 알고리즘을 통해 학습된 신경망으로써 군집하는 과정을 시각화하고, 각 스포츠 종목의 특성을 분석하여 군집하기 위한 도구이다. 본 연구에서는 19개 스포츠 종목의 한국 남자 국가대표 선수 288명을 연구대상으로 선정하였다. 선정된 연구대상은 신장, 좌고, 몸무게, 흉위, 체지방률, 팔굽혀펴기, 농구공던지기, 윗몸일으키기, 하프스쿼트점프, 제자리멀리뛰기, 오래달리기(1600m), 오래달리기(1000m), 50m 달리기, 사이드스텝, 윗몸앞으로굽히기의 체격 및 체력검사를 실시하였다. 수집된 자료는 자기구성지도 툴박스(Self-Organized Map Toolbox)가 탑재된 MATLAB 14.0 프로그램을 이용하여 분석되었는데, 경쟁학습 알고리즘을 통해 설계된 자기구성지도의 크기는 [12 7]로 나타났다. 설계된 자기구성지도는 원자료의 투영 결과와 군집화 기능(Clustering Function)에 따라 6개의 군집을 나타냈다. 군집1에는 사이클, 유도, 탁구, 태권도, 테니스, 군집2에는 육상, 펜싱, 핸드볼, 군집3에는 농구, 배구, 야구, 군집4에는 수영, 양궁, 역도, 군집5에는 럭비, 군집6에는 근대5종, 레슬링, 축구, 하키로 각각 나타났다. 군집 1은 체격·체력요인의 Z-점수가 적게 나타나 체격이 작은 집단으로 구분되었으며, 군집 2는 작지만 민첩성이 좋은 집단으로 나타났다. 군집 3은 신장, 체지방률, 그리고 상지순발력(농구공던지기)와 하지순발력(제자리멀리뛰기)이 좋은 집단으로 나타났으며, 군집 4는 신체구성이 작지만, 심폐지구력이 좋은 것으로 나타났다. 군집 5는 몸무게, 흉위가 다른 군집에 비해 월등히 높게 나타났으며, 군집 6은 하지근지구력(하프스쿼트점프)와 복근지구력(윗몸일으키기)이 좋은 집단으로 나타났다. 자기구성지도는 자료의 처리과정을 시각화 할 수 있는 장점을 지니고 있었으며, 대용량의 자료를 변인의 특성에 맞춰 구분할 수 있어서 자료의 요약과 특성을 분석할 수 있었다. The purpose of current study was to identify the possibility of use the Self-Organized Map among the artificial intelligent techniques into the field of sport science. It is not only the main purposes of the study that the cluster analysis of sport events based on the fitness features was also concerned. The differences of fitness characteristics were utilized with the Self-Organized Map that it was trained based on the competitive learning algorithm. In the study, there were 288 international level athletes who have performing 19 different sport selected as subjects. The Standing Height, Sitting Height, Weight, Girth of Chest and Body Fat % as body measurement and the Push-up, Basketball throw, sit-up, Half-squat Jump, Stand run jump, long run (1600m), run 1000m, run 50m, side step and trunk forward flexion as the fitness were measured. The MATLAB 14.0 package with SOM (Self-Organized Map) toolbox was utilized to cluster the sports, and then the results were discussed within the literatures. As the results of study, the clustering by Self-Organized Map shown 6 clusters. The Cycling, Judo, Table Tennis, Taekwando and Tennis were involved in Cluster 1, Athletics, Fencing and Handball were in Cluster 2, Basketball, Volleyball and Baseball were involved in Cluster 3. The Swimming, Archery and Weight-lifting were in Cluster 4 and Rugby were in Cluster 5. Finally, the modern pentathlon, wrestling, soccer and field hockey were involved in Cluster 6.

Validity Study of Kohonen Self-Organizing Maps

Huh, Myung-Hoe The Korean Statistical Society 2003 Communications for statistical applications and me Vol.10 No.2

Self-organizing map (SOM) has been developed mainly by T. Kohonen and his colleagues as a unsupervised learning neural network. Because of its topological ordering property, SOM is known to be very useful in pattern recognition and text information retrieval areas. Recently, data miners use Kohonen´s mapping method frequently in exploratory analyses of large data sets. One problem facing SOM builder is that there exists no sensible criterion for evaluating goodness-of-fit of the map at hand. In this short communication, we propose valid evaluation procedures for the Kohonen SOM of any size. The methods can be used in selecting the best map among several candidates.

Validity Study of Kohonen Self-Organizing Maps

Myung Hoe Huh 한국통계학회 2003 Communications for statistical applications and me Vol.10 No.2

Self-organizing map (SOM) has been developed mainly by T. Kohonen and his colleagues as a unsupervised learning neural network. Because of its topological ordering property, SOM is known to be very useful in pattern recognition and text information retrieval areas. Recently, data miners use Kohonen`s mapping method frequently in exploratory analyses of large data sets. One problem facing SOM builder is that there exists no sensible criterion for evaluating goodness-of-fit of the map at hand. In this short communication, we propose valid evaluation procedures for the Kohonen SOM of any size. The methods can be used in selecting the best map among several candidates.

Implementation of Machine Part Cell Formation Algorithm in Cellular Manufacturing Technology Using Neural Networks

Tribikram Pradhan,Satya Ranjan Mishra 보안공학연구지원센터 2015 International Journal of Hybrid Information Techno Vol.8 No.2

The machine-part cell formation problem explains the topics surrounding the creation of part families based on the processing needs of the components, and the classification of machine groups based on their capacity to process exact part families in manufacturing. Allocate the specified machines and parts into cells so that the grouping efficiency is maximized. The objective of the paper is to cluster the parts based upon self-organizing map (SOM). After clustering, Minkowski distance is used to sort the machines and some bottlenecks are introduced during the clustering process which can be removed manually by rearranging. We are going to implement self-organizing map (SOM) which is a type of artificial neural network technique to solve this problem. The proposed Minkowski distance approach produced solution with a good group efficiency.

Development of an Application for Mobile Devices to Analyze Data Set by a Self-Organizing Map

Hiroshi Wakuya,Yu Horinouchi,Hideaki Itoh 한국콘텐츠학회(IJOC) 2013 International Journal of Contents Vol.9 No.3

In the preceding studies, an analysis of Saga Prefectural sightseeing information by a Self-Organizing Map (SOM) has been tried. And recent development on Information and Communication Technology (ICT) will help us to access any results via the mobile devices easily. This is why the mobile devices, e.g., smartphones and tablet computers, have an operating system installed, and we can improve their functions by downloading any applications on the Web. Then, in order to realize this basic idea, development of an application for the mobile devices is investigated through some computer simulations on the standard desktop PC in this paper. As a result, it is found that i) a developed feature map is useful to identify some candidate topics, ii) a touchscreen is suitable to show the feature map, and iii) arrangement of the feature map can be modified based on our interests. Then, it is concluded that the proposed idea seems to be applicable, even though further consideration is required to brush it up.