• 제목/요약/키워드: F-Measure

검색결과 1,386건 처리시간 0.031초

정의형 질의응답 시스템을 위한 정답 패턴 (Answer Pattern for Definitional Question-Answering System)

  • 서영훈;신승은
    • 한국콘텐츠학회논문지
    • /
    • 제5권3호
    • /
    • pp.209-215
    • /
    • 2005
  • 본 논문에서는 정의형 질의응답 시스템을 위한 정답 패턴에 대하여 기술한다. 정의형 질의응답 시스템은 정의형 질의에 대한 정답으로 단답형 정답이 아닌 서술형 정답을 제공하기 때문에, 정답 추출 방법이 일반적인 단답형 정답 추출 방법과 다르다. 정의형 정답 패턴을 이용한 정의형 정답 추출은 의미 분석없이 정확한 정의형 정답을 추출할 수 있다. 정의형 정답 패턴은 정확한 정답 추출을 위해 정답 패턴과 패턴별 제약 규칙, 우선순위로 구성된다. 정의형 정답 학습 코퍼스로부터 정답 패턴을 추출하고, 각각의 정답 패턴에 대한F-measure에 따라 최적화하여 패턴별 제약 규칙을 구성한다. 마지막으로 정확률과 정답 패턴 구문 구조를 이용하여 우선순위를 결정한다. 제안한 정의형 정답 패턴을 이용한 정의형 정답 추출은 실험 코퍼스에 대해 정확률 0.8207, 재현율 0.9268, F-measure 0.8705를 보였다. 이것은 제안한 방법이 정의형 질의응답 시스템에 효율적으로 사용될 수 있음을 의미한다.

  • PDF

상황 지식을 이용한 비계층적 군집 기반 하이브리드 추천 (Non-hierarchical Clustering based Hybrid Recommendation using Context Knowledge)

  • 백지원;김민정;박찬홍;정호일;정경용
    • 융합신호처리학회논문지
    • /
    • 제20권3호
    • /
    • pp.138-144
    • /
    • 2019
  • 현대 사회에서 사람들은 시간적인 여유, 경제적인 문제 등에 따라 여행지에 대해 심각한 고민을 한다. 따라서 본 논문에서는 상황 지식을 이용한 비계층적 군집 기반 하이브리드 추천을 제안한다. 제안하는 방법은 사용자의 위치, 장소, 날씨 등의 상황에 따라 선호하는 여행지에 대한 지식을 추천받을 수 있는 개인화된 방법이다. 설문조사를 통해 수집된 데이터로부터 14개의 속성을 기반으로 유사한 특성을 가진 사용자들을 비계층적 군집 기반 하이브리드 추천을 이용하여 군집한다. 이는 암묵적 데이터와 명시적 데이터에 가중치를 부여하여 보다 정확한 추천을 한다. 이를 통해 사용자는 불필요한 시간을 소모하지 않고 선호하는 여행지를 추천받을 수 있다. 성능평가는 정확도, 재현율, F-measure를 이용한다. 평가 결과 정확도는 0.636, 재현율은 0.723, F-measure는 0.676으로 평가되었다.

색상 대비와 텍스처 정보를 이용한 효과적인 스테레오 영상 중요도 맵 추출 (Extraction of an Effective Saliency Map for Stereoscopic Images using Texture Information and Color Contrast)

  • 김성현;강행봉
    • 한국멀티미디어학회논문지
    • /
    • 제18권9호
    • /
    • pp.1008-1018
    • /
    • 2015
  • In this paper, we propose a method that constructs a saliency map in which important regions are accurately specified and the colors of the regions are less influenced by the similar surrounding colors. Our method utilizes LBP(Local Binary Pattern) histogram information to compare and analyze texture information of surrounding regions in order to reduce the effect of color information. We extract the saliency of stereoscopic images by integrating a 2D saliency map with depth information of stereoscopic images. We then measure the distance between two different sizes of the LBP histograms that are generated from pixels. The distance we measure is texture difference between the surrounding regions. We then assign a saliency value according to the distance in LBP histogram. To evaluate our experimental results, we measure the F-measure compared to ground-truth by thresholding a saliency map at 0.8. The average F-Measure is 0.65 and our experimental results show improved performance in comparison with existing other saliency map extraction methods.

피사계심도 확장을 위한 대역통과 필터 기반 초점 정량화 기법 (Bandpass Filter Based Focus Measure for Extended Depth of Field)

  • 차수람;김정태
    • 방송공학회논문지
    • /
    • 제16권5호
    • /
    • pp.883-893
    • /
    • 2011
  • 본 논문에서는, 피사계심도가 낮은 카메라로 획득한 영상에서 초점이 맞는 부분과 그렇지 않은 부분을 나누는 새로운 초점 정량화를 설계하고 초점 정량화에 기반한 decision map을 이용하여 획득한 영상과 위너필터링으로 복원한 영상의 정합을 통해 영상의 피사계심도를 확장하였다. 기존의 초점 정량화 방법은 고주파 성분의 크기에 따라서 초점 정량화를 수행하므로 초점이 맞지 않은 부분에 고주파 성분이 있는 경우 성능이 저하된다는 단점이 있다. 본 논문에서는 이를 극복하기 위하여 효과적인 대역필터를 설계하였으며 제안하는 방법이 기존에 방법에 비해서 성능이 우수함을 시뮬레이션 및 실험을 통해 입증하였다.

뇌졸중 입원 환자의 신체장애 부위별 MMPI 조사연구 (Studies of MMPI on the Somatic Disorder of CVA Patients)

  • 양희숙;장현호;안대종;강형원;류영수;민상준
    • 동의신경정신과학회지
    • /
    • 제13권2호
    • /
    • pp.75-92
    • /
    • 2002
  • To evaluate the relations between strokes and psychological characteristics, the authors surveyed Minnesota Multiphasic Personal Inventory(MMPI) in CVA Patients. The results are as follows.1. In the cerebral infarction patients group, the measure of F, Hy, Pd, Pt and Sc has appeared significantly high. 2. In the cerebral hemorrhage patients group, the measure of F and Ma has appeared significantly high. 3. In dysphasia-hemiparesis patients group, the measure of D has appeared high.4. In the Soyangin patients group, the measure of Pd and Ma has appeared high. 5. In the Soeumin patients group, the measure of D, Pt and Si has appeared high.

  • PDF

급속 동결 융해에 대한 콘크리트의 저항 시험방법(KS F 2456)에 전단파 적용을 위한 실험적 연구 (A Experimental Study on Application of KS F 2456 using Shear Wave)

  • 안지환;전성일;권수안
    • 한국도로학회논문집
    • /
    • 제14권5호
    • /
    • pp.57-65
    • /
    • 2012
  • PURPOSES: It is important to consider the long-term performance of concrete pavement, because concrete pavement is more exposed to the various environmental conditions than any other concrete structures. One of the several methods to evaluate the long-term performance of concrete during winter is KS F 2456. Relative dynamic modulus of elasticity shows the resistance to freezing and thawing. METHODS: To measure relative dynamic modulus of elasticity, ultra sonic is generally used. But in this study, to measure the relative dynamic modulus of elasticity, both ultra sonic and shear wave were used and then compared each other. RESULTS: The results from the measurement by ultrasonic wave and shear wave were divided into three types. Type 1 : Specimens are good and relative dynamic modulus of elasticity did not decrease until 300 cycle. Type 2 : The relative dynamic modulus of elasticity decreased from the late cycle.(about 150 cycle later) Type 3 : The relative dynamic modulus of elasticity consistently decreased from the beginning. As a result of ANOVA, there is no difference according to measuring method, in type 2 and 3. But there is a difference according to measuring method, in type 1's relative dynamic modulus of elasticity. CONCLUSIONS: It is proved that shear wave can be used to understand the damage tendency of relative freezing and thawing and to measure the relative dynamic modulus of elasticity.

키넥트 센서와 유니티 3D 엔진기반의 객체 인식 기법을 적용한 체험형 게임 콘텐츠 설계 및 구현 (A Design and Implementation of Object Recognition based Interactive Game Contents using Kinect Sensor and Unity 3D Engine)

  • 정세훈;이주환;조경호;박재성;심춘보
    • 한국멀티미디어학회논문지
    • /
    • 제21권12호
    • /
    • pp.1493-1503
    • /
    • 2018
  • We propose an object recognition system and experiential game contents using Kinect to maximize object recognition rate by utilizing underwater robots. we implement an ice hockey game based on object-aware interactive contents to validate the excellence of the proposed system. The object recognition system, which is a preprocessor module, is composed based on Kinect and OpenCV. Network sockets are utilized for object recognition communications between C/S. The problem of existing research, degradation of object recognition at long distance, is solved by combining the system development method suggested in the study. As a result of the performance evaluation, the underwater robot object recognized all target objects (90.49%) with 80% of accuracy from a 2m distance, revealing 42.46% of F-Measure. From a 2.5m distance, it recognized 82.87% of the target objects with 60.5% of accuracy, showing 34.96% of F-Measure. Finally, it recognized 98.50% of target objects with 59.4% of accuracy from a 3m distance, showing 37.04% of F-measure.

The Effect of Laughter Therapy on Anxiety, Depression and Stress of Cancer Patients Admitted to Long-Term Care Unit

  • Kim, Seongeun;Kim, Miseon;Jeong, Hyeoncheol
    • Physical Therapy Rehabilitation Science
    • /
    • 제10권2호
    • /
    • pp.197-204
    • /
    • 2021
  • Objective: This study is based on a parallax design before and after a non-equivalent control group to examine the effects of laughter therapy on the anxiety, depression, and stress in patients with cancer admitted to a nursing hospital. Design: A randomized controlled trial. Methods: Data collection was conducted on 34 patients who agreed to participate in the study among patients admitted to E Cancer Hospital from September to November 2018. Laughter treatment interventions were performed three times a week for 50 minutes each and a total of eight times. In this study, a tool developed by Spielberger was used to measure state anxiety, a tool developed by Beck was used to measure depression, and the Perceived Stress Scaled eveloped by Cohen was used to measure stress. Results: Anxiety was significantly lower in the experimental group than in the control group (F=10.74, p=0.003). Depression was lower in both the experimental group and the control group that did not participate in the laughter therapy intervention and was not significant (F=0.58, p=0.451). Stress was significantly lower in the experimental group than in the control group (F=4.36, p=0.045). Conclusions: The results of this study suggest that laughter therapy has a positive effect on reducing anxiety and stress in patients with cancer admitted to nursing hospitals.

뇌파 상대 스펙트럼의 숙련도 평가 척도로의 이용 가능성에 대한 연구 (A Study on Applicability of EEG Spectral Relative Power as a Measure of Expertise Level)

  • 옥동민;박희석
    • 대한인간공학회지
    • /
    • 제29권5호
    • /
    • pp.741-750
    • /
    • 2010
  • The objective of this paper is to study if the EEG spectral relative power would be a reasonable measure of expertise level. EEG electrodes were placed on the locations of Fp1, Fp2, F3, F4, T3, T4, O1, O2 while 5 subjects were playing 4 kinds of game on PC. EEG spectral relative power was significantly related with expertise level on the locations of Fp1, T3, T4, O1, O2. And the results showed that the $\theta$ and $\alpha$ activities were decreased, while $\beta$ and $\gamma$ activities were increased. The results indicated that the EEG spectral relative power would be applicable as a quantitative measure of expertise level.