• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.25 no.3B
• /
• pp.565-569
• /
• 2000

We compose the two-stage recurrent neural network that returns both signals of a hidden and an output layer to the hidden layer. It is tested on the basis of syllables for Korean spoken digit from /gong/to /gu. For these experiments, we adjust the neuron number of the hidden layer, the predictive order of input data and self-recurrent coefficient of the decision state layer. By the experimental results, the recognition rate of this neural network is between 91% and 97.5% in the speaker-dependent case and between 80.75% and 92% in the speaker-independent case. In the speaker-dependent case, this network shows an equivalent recognition performance to Jordan and Elman network but in the speaker-independent case, it does improved performance.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.7 no.2
• /
• pp.182-187
• /
• 2006

This paper proposes a method for boundary recognition of named entity using hidden markov model and ontology information of biological named entity. We uses smoothing method using 31 feature information of word and hierarchical information to alleviate sparse data problem in HMM. The GENIA corpus version 2.1 was used to train and to experiment the proposed boundary recognition system. The experimental results show that the proposed system outperform the previous system which did not use ontology information of hierarchical information and smoothing technique. Also the system shows improvement of execution time of boundary recognition.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.15 no.3
• /
• pp.145-152
• /
• 2015

We consider an online selective-sample learning problem for sequence classification, where the goal is to learn a predictive model using a stream of data samples whose class labels can be selectively queried by the algorithm. Given that there is a limit to the total number of queries permitted, the key issue is choosing the most informative and salient samples for their class labels to be queried. Recently, several aggressive selective-sample algorithms have been proposed under a linear model for static (non-sequential) binary classification. We extend the idea to hidden Markov models for multi-class sequence classification by introducing reasonable measures for the novelty and prediction confidence of the incoming sample with respect to the current model, on which the query decision is based. For several sequence classification datasets/tasks in online learning setups, we demonstrate the effectiveness of the proposed approach.

• The Journal of the Acoustical Society of Korea
• /
• v.10 no.2
• /
• pp.29-35
• /
• 1991

In this paper, we discuss the method of recognizing allophone unit isolated words using hidden Markov model(HMM). Frist we constructed allophone lexicon by extracting allophones from training data and by training allophone HMMs. And then to recognize isolated words using allophone HMMs, it is necessary to construct word dictionary which contains information of allophone sequence and inter-allophone transition probability. Allophone sequences are represented by allophone HMMs. To see the effects of inter-allophone transition probability and to determine optimal probabilities, we performend some experiments. And we showed that small number of traing data and simple train procedure is needed to train word HMMs of allophone sequences and that not less performance than word unit HMM is obtained.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.7 no.11
• /
• pp.2561-2576
• /
• 2013

The ship ad-hoc network (SANET) extends the coverage of the high data-rate terrestrial communications to the ships with the reduced cost in maritime communications. Cognitive radio (CR) has the ability of sensing the radio environment and dynamically reconfiguring the operating parameters, which can make SANET utilize the spectrum efficiently. However, due to the dynamic topology nature and no central entity for data fusion in SANET, the interference brought into the primary network caused by the hidden primary user requires to be carefully managed by a sort of decentralized cooperative spectrum sensing schemes. In this paper, we propose a self-weighted decentralized cooperative spectrum sensing (SWDCSS) scheme to solve such a problem. The analytical and simulation results show that the proposed SWDCSS scheme is reliable to detect the primary user in SANET. As a result, secondary network can efficiently utilize the spectrum band of primary network with little interference to primary network. Referring the complementary receiver operating characteristic (ROC) curves, we observe that with a given false alarm probability, our proposed algorithm reduces the missing probability by 27% than the traditional embedded spectrally agile radio protocol for evacuation (ESCAPE) algorithm in the best condition.

• Proceedings of the Korean Information Science Society Conference
• /
• 2003.04c
• /
• pp.178-180
• /
• 2003

EEG-based brain computer interface(BCI) provides a new communication channel between human brain and computer. EEG data is a multivariate time series so that hidden Markov model (HMM) might be a good choice for classification. However EEG is very noisy data and contains artifacts, so useful features mr expected to improve the performance of HMM. In this paper we addresses the usefulness of principal component features with Hidden Markov model (HHM). We show that some selected principal component features can suppress small noises and artifacts, hence improves classification performance. Experimental study for the classification of EEG data during imagination of a left, right up or down hand movement confirms the validity of our proposed method.

• Magazine of the Korean Society of Agricultural Engineers
• /
• v.43 no.2
• /
• pp.69-77
• /
• 2001

An artificial neural network model was developed to analyze and forecast daily steamflow flow a small watershed. Error Back propagation neural networks (EBPN) of daily rainfall and runoff data were found to have a high performance in simulating stremflow. The model adopts a gradient descent method where the momentum and adaptive learning rate concepts were employed to minimize local minima value problems and speed up the convergence of EBP method. The number of hidden nodes was optimized using Bayesian information criterion. The resulting optimal EBPN model for forecasting daily streamflow consists of three rainfall and four runoff data (Model34), and the best number of the hidden nodes were found to be 13. The proposed model simulates the daily streamflow satisfactorily by comparison compared to the observed data at the HS#3 watershed of the Baran watershed project, which is 391.8 ha and has relatively steep topography and complex land use.

• Speech Sciences
• /
• v.13 no.3
• /
• pp.7-18
• /
• 2006

Diagnosis of pathological voice is one of the important issues in biomedical applications of speech technology. This study focuses on the discrimination of voice disorder using HMM (Hidden Markov Model) for automatic detection between normal voice and vocal fold disorder voice. This is a non-intrusive, non-expensive and fully automated method using only a speech sample of the subject. Speech data from normal people and patients were collected. Mel-frequency filter cepstral coefficients (MFCCs) were modeled by HMM classifier. Different states (3 states, 5 states and 7 states), 3 mixtures and left to right HMMs were formed. This method gives an accuracy of 93.8% for train data and 91.7% for test data in the discrimination of normal and vocal fold disorder voice for sustained /a/.

• Journal of Internet Computing and Services
• /
• v.8 no.5
• /
• pp.141-150
• /
• 2007

This paper proposes new Steganographic model for VoIP that has very effective method using low bit encoding. Most of Steganographic models using Low bit Encoding have two disadvantages; one is that the existence of hidden secret message can be easily detected by auditory, the other is that the capacity of stego data is low. To solve these problems, this method embed more than one bit in inaudible range, so this method can improve the capacity of the hidden message in cover data. The embedding bit position is determined by using a pseudo random number generator which has seed with remaining message length, so it is hard to detect the stego data produced by the proposed method. This proposed model is able to use not only to communicate wave file with hidden message in VoIP environment but also to hide vary information which is user basic information, authentication system, etc.

• Journal of Korea Multimedia Society
• /
• v.21 no.2
• /
• pp.199-209
• /
• 2018

Recently, services provided to consumers are increasingly being combined with big data such as low-priced shopping, customized advertisement, and product recommendation. With the increasing importance of big data, the web crawler that collects data from the web has also become important. However, there are two problems with existing web crawlers. First, if the URL is hidden from the link, it can not be accessed by the URL. The second is the inefficiency of fetching more data than the user wants. Therefore, in this paper, through the Casper.js which can control the DOM in the headless brwoser, DOM event is generated by accessing the URL to the hidden link. We also propose an intelligent web crawler system that allows users to make steps to fine-tune both Structured and unstructured data to bring only the data they want. Finally, we show the superiority of the proposed crawler system through the performance evaluation results of the existing web crawler and the proposed web crawler.