• Journal of the Institute of Convergence Signal Processing
• /
• v.2 no.4
• /
• pp.9-15
• /
• 2001

In this paper we investigate context dependent acoustic models to improve the performance of the Korean speech recognition . The algorithm are using the Korean phonological rules and decision tree, By Successive State Splitting(SSS) algorithm the Hidden Merkov Netwwork(HM-Net) which is an efficient representation of phoneme-context-dependent HMMs, can be generated automatically SSS is powerful technique to design topologies of tied-state HMMs but it doesn't treat unknown contexts in the training phoneme contexts environment adequately In addition it has some problem in the procedure of the contextual domain. In this paper we adopt a new state-clustering algorithm of SSS, called Phonetic Decision Tree-based SSS (PDT-SSS) which includes contexts splits based on the Korean phonological rules. This method combines advantages of both the decision tree clustering and SSS, and can generated highly accurate HM-Net that can express any contexts To verify the effectiveness of the adopted methods. the experiments are carried out using KLE 452 word database and YNU 200 sentence database. Through the Korean phoneme word and sentence recognition experiments. we proved that the new state-clustering algorithm produce better phoneme, word and continuous speech recognition accuracy than the conventional HMMs.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.47 no.1
• /
• pp.104-115
• /
• 2010

Development of the system which provides services using diverse sensors is expanding due to the widespread use of ubiquitous technology, and the research on the security technologies gaining attention to solve the vulnerability of ubiquitous environment's security. However, there are many instances in which flexible security services should be considered instead of strong only security function depending on the context. This paper used Fuzzy algorithm and MAUT to be aware of the diverse contexts and to propose context-aware security service which provides flexible security function according to the context.

• KIPS Transactions on Software and Data Engineering
• /
• v.4 no.9
• /
• pp.409-418
• /
• 2015

In order to delivery of the correct information in IoT environment, it is important to deduce collected information according to a user's situation and to create a new information. In this paper, we propose a control access scheme of information through context-aware to protect sensitive information in IoT environment. It focuses on the access rights management to grant access in consideration of the user's situation, and constrains(access control policy) the access of the data stored in network of unauthorized users. To this end, after analysis of the existing research 'CP-ABE-based on context information access control scheme', then include dynamic conditions in the range of status information, finally we propose a access control policy reflecting the extended multi-dimensional context attribute. Proposed in this paper, access control policy considering the dynamic conditions is designed to suit for IoT sensor fusion environment. Therefore, comparing the existing studies, there are advantages it make a possible to ensure the variety and accuracy of data, and to extend the existing context properties.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.36 no.4B
• /
• pp.354-361
• /
• 2011

Increasing interest and technological advances in smart home has led to active research on context-awareness service and prediction algorithms such as Bayesian Networks, Tree-Dimensional Structures and Genetic prediction algorithms. Context-awareness service presents that providing automatic customized service regarding individual user's pattern surely helps users improve the quality of life. However, it is difficult to implement context-awareness service because the problems are that handling coincidence with context information and exceptional cases have to consider. To overcome this problem, we proposes an Intelligent Sequential Matching Algorithm(ISMA), models context-awareness service using Timed Petri-net(TPN) which is petri-net to have time factor. The example scenario illustrates the effectiveness of the Timed Petri-net model and our proposed algorithm improves average 4~6% than traditional in the accuracy and reliability of prediction.

• The KIPS Transactions:PartD
• /
• v.11D no.7 s.96
• /
• pp.1417-1426
• /
• 2004

As Ubiquitous commerce is coming personalization service is getting interested. And also the recommendation method which offers useful information to customer becomes more important. However, most of them depend on specific method and are restricted to the E-commerce. For applying these recommendation methods into U-commerce, first it is necessary that the extended context modeling and systematic connection of the methods to complement strength and weakness of recommendation methods in each commercial transaction. Therefore, we propose a mod-eling technique of context information related to personal activation in commercial transaction and show incremental preference analysis method, using preference tree which is closely connected to recommendation method in each step. And also, we use an XML indexing technique to effi-ciently extract the recommendation information from a preference tree.

• Journal of Korea Multimedia Society
• /
• v.16 no.7
• /
• pp.810-819
• /
• 2013

FCM (Fuzzy C-Means) clustering algorithm, a typical split-based clustering algorithm, has been successfully applied to the various fields. Nonetheless, the FCM clustering algorithm has some problems, such as high sensitivity to noise and local data, the different clustering result from the intuitive grasp, and the setting of initial round and the number of clusters. To address these problems, in this paper, we determine fuzzy numbers which project the FCM clustering result on the axis with the specific attribute. And we propose a model that the fuzzy numbers apply to FDT (Fuzzy Decision Tree). This model improves the two problems of FCM clustering algorithm such as elevated sensitivity to data, and the difference of the clustering result from the intuitional decision. And also, this paper compares the effect of the proposed model and the result of FCM clustering algorithm through the experiment using real traffic and rainfall data. The experimental results indicate that the proposed model provides more reliable results by the sensitivity relief for data. And we can see that it has improved on the concordance of FCM clustering result with the intuitive expectation.

• The KIPS Transactions:PartD
• /
• v.15D no.1
• /
• pp.23-30
• /
• 2008

The basic research of context aware in ubiquitous environment is an internet technique and XML. The XML data of continuous stream type are popular in network application through the internet. And also there are researches related to query processing for streaming XML data. As a basic research to efficiently query, we propose not only a labeled ordered tree model representing the XML but also a mining method to extract frequent structures from streaming XML data. That is, XML data to continuously be input are modeled by a stream tree which is called by XFP_tree and we exactly extract the frequent structures from the XFP_tree of current window to mine recent data. The proposed method can be applied to the basis of the query processing and index method for XML stream data.

• Journal of Environmental Science International
• /
• v.25 no.5
• /
• pp.635-644
• /
• 2016

Birds play a main role in the formation and change of forest structures as they are seed-dispersal agents. This study aims to identify birds consuming fruits and seeds of the tree and their associated fruits and seeds on Jeju Island in the context on the forest restoration in Korea. We conducted field surveys twice a month from 2013 to 2015 at nine study sites located across Jeju Island and collected available photographic and observation records. A total of 50 species of birds consuming fruits and seeds of the tree were identified and birds belonging to Bombycillidae, Pycnonotidae, Zosteropidae, Sturnidae and Fringillidae were confirmed as major birds consuming fruits and seeds of the tree. Gulping was the dominant type of feeding as well as crushing, but relationship between the size of a bird and the number of fruit and seed species used by the bird was not significant. We also documented that 118 fruit and seed species were consumed by birds and that shrubby fruits and seeds were more consumed by birds than those of other plant types. The relative consumption rate of fruits and seeds ranged from 0.02 to 0.44, but five species were the most important fruits and seeds for birds. Our finding suggest that avian frugivorous gulpers will benefit the seed dispersal, especially of five fruiting plants, providing useful baseline data for forest restoration and urban park design.

• Proceedings of the Korean Information Science Society Conference
• /
• 2005.11b
• /
• pp.667-669
• /
• 2005

유비쿼터스 시대를 향하여 나아가는 현대 사회에서 사람들을 위한 추천시스템은 필수 불가결한 요소 중의 하나이다. 추천 시스템 중에서 사용자의 성별, 나이, 직업 등의 인구 통계적 요소를 고려한 시스템이 주를 이루고 있지만 이러한 시스템에는 어느 정도의 한계가 있다. 추천에 있어서 사용자의 기분, 날씨, 온도 등 주변 환경의 상황이 반영되지 않고 있고 학습을 위한 데이터에 대한 신뢰도 또한 문제가 된다. 이러한 문제점을 해결하기 위해 본 논문에서는 상황정보(Context Information)와 공간 데이터 마이닝(Spatial Data Mining) 기법을 이용한 향상된 추천 시스템을 제안한다. 제안하는 시스템에서는 보다 정확한 추천을 위해 첫째, 날씨, 온도, 사용자의 기분 등의 상황정보를 고려하였다. 그리고 사용자의 유사도 측정을 통해 학습 데이터의 신뢰도를 향상시켰으며, 셋째, 의사결정 트리(Decision Tree) 기법을 이용하여 추천의 정확도를 높였다. 실험을 통하여 측정한 결과 제안하는 추천시스템이 기존의 인구 통계적 요소만을 고려한 시스템이나 의사결정 트리만을 이용한 시스템보다 향상된 성능을 보였다.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1998.06d
• /
• pp.60-63
• /
• 1998

본 논문에서는 사용자가 별도의 훈련과정 없이 인식대상 어휘를 추가 및 변경이 가능한 가변어휘 인식시스템에 관하여 기술한다. 가변어휘 음성인식에서는 미리 구성된 음소모델을 토대로 인식대상 어휘가 결정되명 발음사전에 의거하여 이들 어휘에 해당하는 음소모델을 연결함으로써 단어모델을 만든다. 사용된 음소모델은 현재 음소의 앞뒤의 음소 context를 고려한 문맥종속형(Context-Dependent)음소모델인 triphone을 사용하였고, 연속확률분포를 가지는 Hidden Markov Model(HMM)기반의 고립단어인식 시스템을 구현하였다. 비교를 위해 문맥 독립형 음소모델인 monophone으로 인식실험을 병행하였다. 개발된 시스템은 음성특징벡터로 MFCC(Mel Frequency Cepstrum Coefficient)를 사용하였으며, test 환경에서 나타나지 않은 unseen triphone 문제를 해결하기 위하여 state-tying 방법중 음성학적 지식에 기반을 둔 tree-based clustering 기법을 도입하였다. 음소모델 훈련에는 ETRI에서 구축한 POW (Phonetically Optimized Words) 음성 데이터베이스(DB)[1]를 사용하였고, 어휘독립인식실험에는 POW DB와 관련없는 22개의 부서명을 50명이 발음한 총 1.100개의 고립단어 부서 DB[2]를 사용하였다. 인식실험결과 문맥독립형 음소모델이 88.6%를 보인데 비해 문맥종속형 음소모델은 96.2%의 더 나은 성능을 보였다.