• Journal of KIISE
• /
• v.44 no.1
• /
• pp.57-62
• /
• 2017

Traditional Korean morphological analysis and POS tagging methods usually consist of two steps: 1 Generat hypotheses of all possible combinations of morphemes for given input, 2 Perform POS tagging search optimal result. require additional resource dictionaries and step could error to the step. In this paper, we tried to solve this problem end-to-end fashion using sequence-to-sequence model convolutional features. Experiment results Sejong corpus sour approach achieved 97.15% F1-score on morpheme level, 95.33% and 60.62% precision on word and sentence level, respectively; s96.91% F1-score on morpheme level, 95.40% and 60.62% precision on word and sentence level, respectively.

• The KIPS Transactions:PartB
• /
• v.19B no.3
• /
• pp.195-200
• /
• 2012

In Korean, spelling change in various forms must be recovered into base forms in morphological analysis as well as part-of-speech (POS) tagging is difficult without morphological analysis because Korean is agglutinative. This is one of notorious problems in Korean morphological analysis and has been solved by morpheme recovery rules, which generate morphological ambiguity resolved by POS tagging. In this paper, we propose a morpheme recovery scheme based on machine learning methods like Na$\ddot{i}$ve Bayes models. Input features of the models are the surrounding context of the syllable which the spelling change is occurred and categories of the models are the recovered syllables. The POS tagging system with the proposed model has demonstrated the $F_1$-score of 97.5% for the ETRI tree-tagged corpus. Thus it can be decided that the proposed model is very useful to handle morpheme recovery in Korean.

• Journal of KIISE:Software and Applications
• /
• v.37 no.12
• /
• pp.946-950
• /
• 2010

Many sentiment categorization systems based on machine learning methods use morphological analyzers in order to extract linguistic features from sentences. However, the morphological analyzers do not generally perform well in a customer review domain because online customer reviews include many spacing errors and spelling errors. These low performances of the underlying systems lead to performance decreases of the sentiment categorization systems. To resolve this problem, we propose a feature extraction method based on simple longest matching of Eojeol (a Korean spacing unit) and phoneme patterns. The two kinds of patterns are automatically constructed from a large amount of POS (part-of-speech) tagged corpus. Eojeol patterns consist of Eojeols including content words such as nouns and verbs. Phoneme patterns consist of leading consonant and vowel pairs of predicate words such as verbs and adjectives because spelling errors seldom occur in leading consonants and vowels. To evaluate the proposed method, we implemented a sentiment categorization system using a SVM (Support Vector Machine) as a machine learner. In the experiment with Korean customer reviews, the sentiment categorization system using the proposed method outperformed that using a morphological analyzer as a feature extractor.

• The Journal of the Acoustical Society of Korea
• /
• v.35 no.6
• /
• pp.501-509
• /
• 2016

We propose a method to improve the robustness of speaker verification on short test utterances. The accuracy of the state-of-the-art i-vector/probabilistic linear discriminant analysis systems can be degraded when testing utterance durations are short. The proposed method compensates for utterance variations of short test feature vectors using deep neural networks. We design three different types of DNN (Deep Neural Network) structures which are trained with different target output vectors. Each DNN is trained to minimize the discrepancy between the feed-forwarded output of a given short utterance feature and its original long utterance feature. We use short 2-10 s condition of the NIST (National Institute of Standards Technology, U.S.) 2008 SRE (Speaker Recognition Evaluation) corpus to evaluate the method. The experimental results show that the proposed method reduces the minimum detection cost relative to the baseline system.

• Journal of Korean language education
• /
• v.28 no.4
• /
• pp.181-219
• /
• 2017

The purpose of this study is to review the functions of {됐어} and {X了} as a discourse marker on different levels, and to examine their similarities and differences. {됐어} has not been widely recognized as a discourse marker in the field of Korean linguistics and Korean language education. Therefore, in order to establish the identity of {됐어} as a discourse marker, the reasons that {됐어} can be regarded as discourse marker were explained prior to the contrastive analysis. As to the method of contrastive analysis for {됐어} and {X了}, they were analyzed on three main dimensions: that is, the textual dimension, the interpersonal dimension, and the metalinguistic dimension in the corpus consisting of scripts of Korean and Chinese sitcoms. The results are as follows. In the textual domain, {됐어} and {X了} have the function of closing the topic in common, while {X了} can indicate a new topic and transmit a topic. In terms of functions in the interpersonal domain, {됐어} and {X了} are commonly used to refuse a partner's proposal or request and to interrupt a partner's speech or action. Furthermore, in the interactional aspect, {됐어} and {X了} performs the function of expressing a response to a preceding utterance and taking the turn of speaking. The difference between them in the interpersonal domain is that {X了} performs the function of correcting a speaker's utterance. In the metalinguistic domain, {됐어} and {X了} are common in that they perform the function of expressing the dissatisfaction of the speaker, showing generosity and making a compromise with the addressee. {X了}'s distinguishing characteristics in this domain is that it can express the attitude of consoling the hearer.

• The Korean Journal of Applied Statistics
• /
• v.34 no.3
• /
• pp.295-308
• /
• 2021

In recent years, as many people post their interests on social media or store documents in digital form due to the development of the internet and computer technologies, the amount of text data generated has exploded. Accordingly, the demand for technology to create valuable information from numerous document data is also increasing. In this study, through statistical techniques, we investigate how the meanings of Korean words change over time by using the presidential speech records and newspaper articles public data. Using this, we present a strategy that can be utilized in the study of the synchronic change of Hangeul. The purpose of this study is to deviate from the study of the theoretical language phenomenon of Hangeul, which was studied by the intuition of existing linguists or native speakers, to derive numerical values through public documents that can be used by anyone, and to explain the phenomenon of changes in the meaning of words.

• The Journal of the Acoustical Society of Korea
• /
• v.41 no.3
• /
• pp.351-358
• /
• 2022

Recently, Generative Adversarial Networks (GAN) and Variational AutoEncoders (VAE) have been applied to voice conversion that can make use of non-parallel training data. Especially, Conditional Cycle-Consistent Generative Adversarial Networks (CC-GAN) and Cycle-Consistent Variational AutoEncoders (CycleVAE) show promising results in many-to-many voice conversion among multiple speakers. However, the number of speakers has been relatively small in the conventional voice conversion studies using the CC-GANs and the CycleVAEs. In this paper, we extend the number of speakers to 100, and analyze the performances of the many-to-many voice conversion methods experimentally. It has been found through the experiments that the CC-GAN shows 4.5 % less Mel-Cepstral Distortion (MCD) for a small number of speakers, whereas the CycleVAE shows 12.7 % less MCD in a limited training time for a large number of speakers.

• Journal of Intelligence and Information Systems
• /
• v.23 no.3
• /
• pp.119-138
• /
• 2017

Recently, SNS has become an important channel for marketing as well as personal communication. However, cybercrime has also evolved with the development of information and communication technology, and illegal advertising is distributed to SNS in large quantity. As a result, personal information is lost and even monetary damages occur more frequently. In this study, we propose a method to analyze which sentences and documents, which have been sent to the SNS, are related to financial fraud. First of all, as a conceptual framework, we developed a matrix of conceptual characteristics of cybercriminality on SNS and emergency management. We also suggested emergency management process which consists of Pre-Cybercriminality (e.g. risk identification) and Post-Cybercriminality steps. Among those we focused on risk identification in this paper. The main process consists of data collection, preprocessing and analysis. First, we selected two words 'daechul(loan)' and 'sachae(private loan)' as seed words and collected data with this word from SNS such as twitter. The collected data are given to the two researchers to decide whether they are related to the cybercriminality, particularly financial fraud, or not. Then we selected some of them as keywords if the vocabularies are related to the nominals and symbols. With the selected keywords, we searched and collected data from web materials such as twitter, news, blog, and more than 820,000 articles collected. The collected articles were refined through preprocessing and made into learning data. The preprocessing process is divided into performing morphological analysis step, removing stop words step, and selecting valid part-of-speech step. In the morphological analysis step, a complex sentence is transformed into some morpheme units to enable mechanical analysis. In the removing stop words step, non-lexical elements such as numbers, punctuation marks, and double spaces are removed from the text. In the step of selecting valid part-of-speech, only two kinds of nouns and symbols are considered. Since nouns could refer to things, the intent of message is expressed better than the other part-of-speech. Moreover, the more illegal the text is, the more frequently symbols are used. The selected data is given 'legal' or 'illegal'. To make the selected data as learning data through the preprocessing process, it is necessary to classify whether each data is legitimate or not. The processed data is then converted into Corpus type and Document-Term Matrix. Finally, the two types of 'legal' and 'illegal' files were mixed and randomly divided into learning data set and test data set. In this study, we set the learning data as 70% and the test data as 30%. SVM was used as the discrimination algorithm. Since SVM requires gamma and cost values as the main parameters, we set gamma as 0.5 and cost as 10, based on the optimal value function. The cost is set higher than general cases. To show the feasibility of the idea proposed in this paper, we compared the proposed method with MLE (Maximum Likelihood Estimation), Term Frequency, and Collective Intelligence method. Overall accuracy and was used as the metric. As a result, the overall accuracy of the proposed method was 92.41% of illegal loan advertisement and 77.75% of illegal visit sales, which is apparently superior to that of the Term Frequency, MLE, etc. Hence, the result suggests that the proposed method is valid and usable practically. In this paper, we propose a framework for crisis management caused by abnormalities of unstructured data sources such as SNS. We hope this study will contribute to the academia by identifying what to consider when applying the SVM-like discrimination algorithm to text analysis. Moreover, the study will also contribute to the practitioners in the field of brand management and opinion mining.

• Korean Journal of Cognitive Science
• /
• v.22 no.1
• /
• pp.57-78
• /
• 2011

Paraphrasing is the act of writing a text using other words without altering the meaning. Paraphrases can be used in many fields of natural language processing. In particular, paraphrases can be incorporated in machine translation in order to improve the coverage and the quality of translation. Recently, the approaches on paraphrase extraction utilize bilingual parallel corpora, which consist of aligned sentence pairs. In these approaches, paraphrases are identified, from the word alignment result, by pivot phrases which are the phrases in one language to which two or more phrases are connected in the other language. However, the word alignment is itself a very difficult task, so there can be many alignment errors. Moreover, the alignment errors can lead to the problem of selecting incorrect pivot phrases. In this study, we propose a method in paraphrase extraction that discriminates good pivot phrases from bad pivot phrases. Each pivot phrase is weighted according to its reliability, which is scored by considering the lexical and part-of-speech information. The experimental result shows that the proposed method achieves higher precision and recall of the paraphrase extraction than the baseline. Also, we show that the extracted paraphrases can increase the coverage of the Korean-English machine translation.

• Journal of The Korean Society of Inherited Metabolic disease
• /
• v.15 no.2
• /
• pp.87-92
• /
• 2015

Mucopolysaccharidosis (MPS) is an inherited disease entity associated with lysosomal enzyme deficiencies. MPS type 2, also known as Hunter syndrome, has a characteristic morphology primarily involving x-l inked recessive defects and iduronate-2-sulfatase gene mutation. The purpose of this case report is to provide important clues to help pediatricians identify Hunter syndrome patients earlier (i.e., before the disease progresses). A 30-month-old boy showed developmental delay and decreased speech ability. Physical examinations revealed a flat nose and extensive Mongolian spots. Brain magnetic resonance images (MRIs) showed bilateral multiple patchy T2 hyperintense lesions in the periventricular and deep white matter, several cyst-like lesions in the body of the corpus callosum, and diffuse brain atrophy, which were in keeping with the diagnosis. Based on these findings, the patient was suspected of having MPS. In the laboratory findings, although the genetic analysis of IDS (Iduronate-2-sulfatase) did not show any pathogenic variant, the enzymatic activity of IDS was not detected. We could confirm the diagnosis of MPS, because other sulfatases, such as ${\alpha}$-L-iduronidase, were detected in the normal range. Early enzymatic replacement therapy is essential and has a relatively good prognosis. Therefore, early diagnosis should be made before organ damage becomes irreversible, and brain MRIs can provide additional diagnostic clues to help distinguish the disorder.