• Title/Summary/Keyword: one-class classification

Search Result 353, Processing Time 0.028 seconds

Weighted Least Squares Based on Feature Transformation using Distance Computation for Binary Classification (이진 분류를 위하여 거리계산을 이용한 특징 변환 기반의 가중된 최소 자승법)

  • Jang, Se-In;Park, Choong-Shik
    • Journal of the Korea Institute of Information and Communication Engineering
    • /
    • v.24 no.2
    • /
    • pp.219-224
    • /
    • 2020
  • Binary classification has been broadly investigated in machine learning. In addition, binary classification can be easily extended to multi class problems. To successfully utilize machine learning methods for classification tasks, preprocessing and feature extraction steps are essential. These are important steps to improve their classification performances. In this paper, we propose a new learning method based on weighted least squares. In the weighted least squares, designing weights has a significant role. Due to this necessity, we also propose a new technique to obtain weights that can achieve feature transformation. Based on this weighting technique, we also propose a method to combine the learning and feature extraction processes together to perform both processes simultaneously in one step. The proposed method shows the promising performance on five UCI machine learning data sets.

Online anomaly detection algorithm based on deep support vector data description using incremental centroid update (점진적 중심 갱신을 이용한 deep support vector data description 기반의 온라인 비정상 탐지 알고리즘)

  • Lee, Kibae;Ko, Guhn Hyeok;Lee, Chong Hyun
    • The Journal of the Acoustical Society of Korea
    • /
    • v.41 no.2
    • /
    • pp.199-209
    • /
    • 2022
  • Typical anomaly detection algorithms are trained by using prior data. Thus the batch learning based algorithms cause inevitable performance degradation when characteristics of newly incoming normal data change over time. We propose an online anomaly detection algorithm which can consider the gradual characteristic changes of incoming normal data. The proposed algorithm based on one-class classification model includes both offline and online learning procedures. In offline learning procedure, the algorithm learns the prior data to be close to centroid of the latent space and then updates the centroid of the latent space incrementally by new incoming data. In the online learning, the algorithm continues learning by using the updated centroid. Through experiments using public underwater acoustic data, the proposed online anomaly detection algorithm takes only approximately 2 % additional learning time for the incremental centroid update and learning. Nevertheless, the proposed algorithm shows 19.10 % improvement in Area Under the receiver operating characteristic Curve (AUC) performance compared to the offline learning model when new incoming normal data comes.

Accuracy of one-step automated orthodontic diagnosis model using a convolutional neural network and lateral cephalogram images with different qualities obtained from nationwide multi-hospitals

  • Yim, Sunjin;Kim, Sungchul;Kim, Inhwan;Park, Jae-Woo;Cho, Jin-Hyoung;Hong, Mihee;Kang, Kyung-Hwa;Kim, Minji;Kim, Su-Jung;Kim, Yoon-Ji;Kim, Young Ho;Lim, Sung-Hoon;Sung, Sang Jin;Kim, Namkug;Baek, Seung-Hak
    • The korean journal of orthodontics
    • /
    • v.52 no.1
    • /
    • pp.3-19
    • /
    • 2022
  • Objective: The purpose of this study was to investigate the accuracy of one-step automated orthodontic diagnosis of skeletodental discrepancies using a convolutional neural network (CNN) and lateral cephalogram images with different qualities from nationwide multi-hospitals. Methods: Among 2,174 lateral cephalograms, 1,993 cephalograms from two hospitals were used for training and internal test sets and 181 cephalograms from eight other hospitals were used for an external test set. They were divided into three classification groups according to anteroposterior skeletal discrepancies (Class I, II, and III), vertical skeletal discrepancies (normodivergent, hypodivergent, and hyperdivergent patterns), and vertical dental discrepancies (normal overbite, deep bite, and open bite) as a gold standard. Pre-trained DenseNet-169 was used as a CNN classifier model. Diagnostic performance was evaluated by receiver operating characteristic (ROC) analysis, t-stochastic neighbor embedding (t-SNE), and gradient-weighted class activation mapping (Grad-CAM). Results: In the ROC analysis, the mean area under the curve and the mean accuracy of all classifications were high with both internal and external test sets (all, > 0.89 and > 0.80). In the t-SNE analysis, our model succeeded in creating good separation between three classification groups. Grad-CAM figures showed differences in the location and size of the focus areas between three classification groups in each diagnosis. Conclusions: Since the accuracy of our model was validated with both internal and external test sets, it shows the possible usefulness of a one-step automated orthodontic diagnosis tool using a CNN model. However, it still needs technical improvement in terms of classifying vertical dental discrepancies.

A Study On Malocclusion Patients From Department Of Orthodontics, Chong-A Dental Hospital (청아치과병원 교정과에 내원한 환자의 분포와 부정교합의 유형)

  • Kim, Nam-Joong;Lee, Chung-Jae
    • Journal of Technologic Dentistry
    • /
    • v.29 no.2
    • /
    • pp.197-211
    • /
    • 2007
  • With the development of orthodontics and increasing concerns on physical appearance, the number of patients has been steadily increasing. It is quite important not only to make effective cure plans and accurate diagnoses but also to have a thorough grasp of patients' malocclusion types and their occurrence frequency, in addition to patients' personality in order to cure the patients appropriately. This study is based on 946 malocclusion patients who had visited Chong-A Dental Hospital from 1999 to 2004 and investigated their aspects of malocclusion and characteristics of their gender, age and residence. The results are as follows. 1. The number of patients per year had been decreased until 2001, after which year the number had fluctuated. The number was the largest in 1999, 169 and the smallest in 2001, 140. Female occupied 68.0% of the total, twice as many as male, 32.0%) 2. Based on the Angle's classification, 19 or over year - old group was the largest of the total, 59.3% and 6 or younger year - old group, the smallest, 0.5%. The 19 or over year old group was less than a half of the total (47.4%) in 2003 and there were no patients who belonged to the 6 or younger year - old group in 2003 and 2004. 3. Distributions on the types of malocclusion have shown that 39.9 % of the total are in the Class I, the largest, 31.0% in the Class I and 29.2 in the Class II, the smallest. 1) The number of the ClassI was 73, the largest, that of the Class III being 35, the smallest in 1999. On the whole, the number of the Class I accounted for the largest part of the total. 2) The number of male patients in the Class II was the smallest, generally being the largest in the Class I. In case of female, that of the Class III was the smallest. 3) Based on the age, the Class I was the highest in between 7 and 13 age group, the Class III the lowest. The Class I occupied the largest around 40%. 4) In the shape of physiognomy, the meso occupied the largest part among all the Class, of which the Class II was the highest, 64.2%. The bracy was the largest in the Class I, and the dolicho in the Class III. 5) In the profile, the convex shape was the largest in the Class I and II, and especially in the Class II, over 3/4 of the total, 75.4%. In contrast, the direct shape was the largest in the Class III and the sunken shape occupied 33.3%, which was nearly ten times more than the case of the Class I and III. 6) In the asymmetry of physiognomy, the number of patients of the Class IIIwas the largest, 34.1% and that of the Class II, the smallest, 19.5%. It was found that about one fourth of the malocclusion patients were under the asymmetry of physiognomy. 4. In the distribution of patients' residence, 81.4% were from the Seoul Metropolis and 48.2% from Gangnam-Gu where Chong-A Dental Hospital is located and Seocho-Gu and Songpa-Gu which are adjacent to Gangnam-Gu.

  • PDF

tufA gene as molecular marker for freshwater Chlorophyceae

  • Vieira, Helena Henriques;Bagatini, Inessa Lacativa;Guinart, Carla Marques;Vieira, Armando Augusto Henriques
    • ALGAE
    • /
    • v.31 no.2
    • /
    • pp.155-165
    • /
    • 2016
  • Green microalgae from the class Chlorophyceae represent a major biodiversity component of eukaryotic algae in continental water. Identification and classification of this group through morphology is a hard task, since it may present cryptic species and phenotypic plasticity. Despite the increasing use of molecular methods for identification of microorganisms, no single standard barcode marker is yet established for this important group of green microalgae. Some available studies present results with a limited number of chlorophycean genera or using markers that require many different primers for different groups within the class. Thus, we aimed to find a single marker easily amplified and with wide coverage within Chlorophyceae using only one pair of primers. Here, we tested the universality of primers for different genes (tufA, ITS, rbcL, and UCP4) in 22 strains, comprising 18 different species from different orders of Chlorophyceae. The ITS primers sequenced only 3 strains and the UCP primer failed to amplify any strain. We tested two pairs of primers for rbcL and the best pair provided sequences for 10 strains whereas the second one provided sequences for only 7 strains. The pair of primers for the tufA gene presented good results for Chlorophyceae, successfully sequencing 21 strains and recovering the expected phylogeny relationships within the class. Thus, the tufA marker stands out as a good choice to be used as molecular marker for the class.

The New Classification of Mountains in the Korean Peninsula and the Mountain Associated Influence on Atmospheric Environment (한반도 산맥의 재조사와 분류 및 대기환경에 미치는 영향)

  • Chung, Yong-Seung;Kim, Hak-Sung
    • Journal of the Korean earth science society
    • /
    • v.37 no.1
    • /
    • pp.21-28
    • /
    • 2016
  • Mountains have significant influences on the atmospheric environment. The Korean Peninsula consists of approximately 70% mountainous area with numerous mountain ranges and peaks. The initial classification of mountains in Korea was made by a Japanese scientist from 1900 to 1902. In fact, the Japanese study created too many names of mountains to maintain, which led to confusions. The purpose of this study aims to simplify the previous names and classification of mountains in the Korean Peninsula so that they can be utilized for educational and general purpose of the society and educational institutions. Through this study, we name various mountains as one name "Korea Mountains" which is classified as the secondary world-mountain class stretching from the Korean Peninsula to northeast China (southern Manchuria). The Korea Mountains connect the third class regional medium-scale mountains of Jirin, Hamkyoung, Taebaek, and the fourth mountain class, 8 small-scale ranges including the Liaoning, Yaenbaen, Hambeuk, Pyoungbeuk, Whanghae, Charyoung, Kyoungsang and Namhae Mountains. The major mountains in the Korean Peninsula are normally influenced by the general circulation of the atmosphere of the world. The atmospheric conditions are modified on the up-stream and down-stream sides; there is a need for continuous monitoring of the atmospheric environment which impacts the ecosystem and human society.

A Measurement of Traffic Vehicles Flow by the Ultrasonic Spatial Filtering Method (교통난 계측 I-초음파용 공간필터법에 의하여-)

  • 전승환
    • Journal of the Korean Institute of Navigation
    • /
    • v.20 no.2
    • /
    • pp.51-58
    • /
    • 1996
  • For the smooth flow of traffic vehicles and its effective management, it is necessary to have an exact information on traffic condition, i.e., the volume of traffic, velocity, occupancy and classification of vehicles. In particular, for classification of vehicles, there has been only image processing method using camera, where the method can obtain much information but rather expensive. In this paper, an algorithm for the measurement of velocity and total length of vehicles has been proposed to develop a general traffic management system, which is necessary to discriminate the class of vehicles. In order to realize the proposed algorithm, we have developed an ultrasonic spatial filtering method, which has better performance than that of using the traditional vehicle detector. To have this system to be constructed, we have introduced three sets of ultrasonic devices where each has one transmitter and two receivers which are arranged to obtain the spatial difference of objects. The velocity of vehicles can be measured by analyzing the occurrence time of pulses and their time differences. The total length of vehicles can be given by multiplying velocity with time interval of pulses sequence. To confirm the effectiveness of this measuring system, the experiment by the spatial filtering method using the ultrasonic sensors has been carried out. As the results, it is found that the proposed method can be used as one of measurement tools in the general traffic management system.

  • PDF

A Study on Novelty Detection of GPS Data Using Human Mobility and OCSVM(One-class SVM) (OCSVM(One-class SVM)과 인간의 이동을 이용한 GPS 데이터의 이상 현상 검출에 관한연구)

  • Kim, Woo-Joong;Song, Ha-Yoon
    • Proceedings of the Korea Information Processing Society Conference
    • /
    • 2011.04a
    • /
    • pp.1060-1063
    • /
    • 2011
  • 인간은 목적지를 향하여 가는 방법의 선택에 있어서 가고자 하는 목적, 목적지, 출발 시간 등에 영향을 받는다. 그러나 이러한 매개변수들과 더불어 중요하게 고려되는 것은 바로 인간의 습관이다. 다시 말해 인간이 목적지로 가는 방법을 선택하는데 습관이라는 매개변수와 밀접한 영향이 있다는 것이다. 이를 미루어 볼 때, 인간의 이동은 습관으로 인해 대부분 특정한 범주 안에서 이동을 할 것이라는 추측할 수 있다. 나아가, 사람들이 흔히 들고 다니는 GPS장치에서 측정된 데이터가 추측한 속성으로 인해 범주를 벗어나는 이상현상을 검출하는 것으로 확장을 할 수 있다. 즉, GPS장치에서 측정된 데이터는 개인별로 클래스화(Classification)가 가능함을 추론할 수 있다. 본 논문에서는 실제 사람이 이동한 좌표를 바탕으로 시간당 변화량을 계산하여 좌표에 사상시켰다. 그리고, 단일 클래스 서포트 백터 머신(OCSVM)을 가지고 클래스화 했으며, OCSVM의 커널 함수 내의 변수인에 따라 클래스의 크기 혹은 클래스 내부의 밀도에 영향을 받음을 알 수 있었으며, 그 둘 사이에는 적절한 교환(Tradeoff)이 발생하였다는 결론이 나왔다.

Simultaneous Optimization of Gene Selection and Tumor Classification Using Intelligent Genetic Algorithm and Support Vector Machine

  • Huang, Hui-Ling;Ho, Shinn-Ying
    • Proceedings of the Korean Society for Bioinformatics Conference
    • /
    • 2005.09a
    • /
    • pp.57-62
    • /
    • 2005
  • Microarray gene expression profiling technology is one of the most important research topics in clinical diagnosis of disease. Given thousands of genes, only a small number of them show strong correlation with a certain phenotype. To identify such an optimal subset from thousands of genes is intractable, which plays a crucial role when classify multiple-class genes express models from tumor samples. This paper proposes an efficient classifier design method to simultaneously select the most relevant genes using an intelligent genetic algorithm (IGA) and design an accurate classifier using Support Vector Machine (SVM). IGA with an intelligent crossover operation based on orthogonal experimental design can efficiently solve large-scale parameter optimization problems. Therefore, the parameters of SVM as well as the binary parameters for gene selection are all encoded in a chromosome to achieve simultaneous optimization of gene selection and the associated SVM for accurate tumor classification. The effectiveness of the proposed method IGA/SVM is evaluated using four benchmark datasets. It is shown by computer simulation that IGA/SVM performs better than the existing method in terms of classification accuracy.

  • PDF

CANCER CLASSIFICATION AND PREDICTION USING MULTIVARIATE ANALYSIS

  • Shon, Ho-Sun;Lee, Heon-Gyu;Ryu, Keun-Ho
    • Proceedings of the KSRS Conference
    • /
    • v.2
    • /
    • pp.706-709
    • /
    • 2006
  • Cancer is one of the major causes of death; however, the survival rate can be increased if discovered at an early stage for timely treatment. According to the statistics of the World Health Organization of 2002, breast cancer was the most prevalent cancer for all cancers occurring in women worldwide, and it account for 16.8% of entire cancers inflicting Korean women today. In order to classify the type of breast cancer whether it is benign or malignant, this study was conducted with the use of the discriminant analysis and the decision tree of data mining with the breast cancer data disclosed on the web. The discriminant analysis is a statistical method to seek certain discriminant criteria and discriminant function to separate the population groups on the basis of observation values obtained from two or more population groups, and use the values obtained to allow the existing observation value to the population group thereto. The decision tree analyzes the record of data collected in the part to show it with the pattern existing in between them, namely, the combination of attribute for the characteristics of each class and make the classification model tree. Through this type of analysis, it may obtain the systematic information on the factors that cause the breast cancer in advance and prevent the risk of recurrence after the surgery.

  • PDF