• MALSORI
• /
• v.68
• /
• pp.115-126
• /
• 2008

In this paper, we propose a vocabulary coverage improvement method for embedded continuous speech recognition (CSR) using knowledgebase. A vocabulary in CSR is normally derived from a word frequency list. Therefore, the vocabulary coverage is dependent on a corpus. In the previous research, we presented an improved way of vocabulary generation using part-of-speech (POS) tagged corpus. We analyzed all words paired with 101 among 152 POS tags and decided on a set of words which have to be included in vocabularies of any size. However, for the other 51 POS tags (e.g. nouns, verbs), the vocabulary inclusion of words paired with such POS tags are still based on word frequency counted on a corpus. In this paper, we propose a corpus independent word inclusion method for noun-, verb-, and named entity(NE)-related POS tags using knowledgebase. For noun-related POS tags, we generate synonym groups and analyze their relative importance using Google search. Then, we categorize verbs by lemma and analyze relative importance of each lemma from a pre-analyzed statistic for verbs. We determine the inclusion order of NEs through Google search. The proposed method shows better coverage for the test short message service (SMS) text corpus.

• Korean Journal of Cognitive Science
• /
• v.19 no.3
• /
• pp.257-281
• /
• 2008

The expression 'kes' is one of the most widely used ones in the language whose uses are highly dependent upon the context. These highly-context dependent uses make it hard to determine its grammatical properties. As a way of examining the properties in a rather controlled context, this paper collects a series of speeches made by government officials and examines the grammatical properties of the expression in the corpus. In particular, the paper, based on the 539 instances of 'kes' uses extracted from the corpus, focuses on the 7 types of 'kes' constructions most widely used in the collected speech corpus.

• Asia Pacific Journal of Corpus Research
• /
• v.3 no.1
• /
• pp.35-49
• /
• 2022

This study examines crosslinguistic influence on the use of concessive conditionals by Japanese EFL learners. Contrastive analysis suggests that Japanese native speakers may overuse the concessive conditional even if due to partial similarities to Japanese concessive conditionals, whose formal and semantic restrictions are fewer than those of English concessive conditionals. This hypothesis is tested using data from the written module of the International Corpus Network of Asian Learners of English (ICNALE). Comparison of Japanese native speakers with English native speakers and Chinese native speakers reveals the following trends. First, Japanese native speakers tend to overuse concessive conditionals compared to native speakers, while similar overuse is not observed in Chinese native speaker data. Second, non-nativelike uses of even if appear in contexts allowing the use of concessive conditionals in Japanese. Third, while overuse and infelicitous use of even if is observed at all proficiency levels, formal errors are restricted to learners at lower proficiency levels. These findings suggest that crosslinguistic influence does occur in the use of concessive conditionals, and that its particular realization is affected by L2 proficiency, with formal crosslinguistic influence mediated at an earlier stage than semantic cross-linguistic influence.

• Proceedings of the KSPS conference
• /
• 2002.11a
• /
• pp.209-213
• /
• 2002

This paper describes the reduction of DB without degradation of speech quality in Corpus-based Speech synthesizer of Korean language. In this paper, it is proposed that the frequency of every unit in reduced DB should reflect the frequency of units in Korean language. So, the target population of every unit is set to be proportional to their frequency in Korean large corpus(780K sentences, 45Mega phonemes). Second, the frequent instances during synthesis should be also maintained in reduced DB. To the last, it is proposed that frequency of every instance should be reflected in clustering criterion and used as criterion for selection of representative instances. The evaluation result with proposed methods reveals better quality than using conventional methods.

• MALSORI
• /
• no.44
• /
• pp.145-156
• /
• 2002

This paper describes the reduction of DB without degradation of speech quality in Corpus-based Speech synthesizer of the Korean language. In this paper, it is proposed that the frequency of every unit in reduced DB reflect the frequency of units in the Korean language. So, the target population of every unit is set to be proportional to its frequency in Korean large corpus (780k sentences, 45Mega phones). Secondly, the frequent instances during synthesis should be also maintained in reduced DB. To the last, it is proposed that frequency of every instance be reflected in clustering criteria and used as another important criterion for selection of representative instances. The evaluation result with proposed methods reveals better quality than that using conventional methods.

• Journal of Korea Multimedia Society
• /
• v.23 no.4
• /
• pp.595-602
• /
• 2020

We propose a generative probabilistic model with Dirichlet prior distribution for topic modeling and text similarity analysis. It assigns a topic and calculates text correlation between documents within a corpus. It also provides posterior probabilities that are assigned to each topic of a document based on the prior distribution in the corpus. We then present a Gibbs sampling algorithm for inference about the posterior distribution and compute text correlation among 50 abstracts from the papers published by IEEE. We also conduct a supervised learning to set a benchmark that justifies the performance of the LDA (Latent Dirichlet Allocation). The experiments show that the accuracy for topic assignment to a certain document is 76% for LDA. The results for supervised learning show the accuracy of 61%, the precision of 93% and the f1-score of 96%. A discussion for experimental results indicates a thorough justification based on probabilities, distributions, evaluation metrics and correlation coefficients with respect to topic assignment.

• Proceedings of the KSPS conference
• /
• 2005.11a
• /
• pp.103-106
• /
• 2005

In this paper, we propose a language modeling approach to improve the performance of a large vocabulary continuous speech recognition system. The proposed approach is based on the active learning framework that helps to select a text corpus from a plenty amount of text data required for language modeling. The perplexity is used as a measure for the corpus selection in the active learning. From the recognition experiments on the task of continuous Korean speech, the speech recognition system employing the language model by the proposed language modeling approach reduces the word error rate by about 6.6 % with less computational complexity than that using a language model constructed with randomly selected texts.

• The KIPS Transactions:PartB
• /
• v.16B no.5
• /
• pp.427-434
• /
• 2009

With the rapid evolution of the Internet and mobile environments, text including spelling errors such as newly-coined words and abbreviated words are widely used. These spelling errors make it difficult to develop NLP (natural language processing) applications because they decrease the readability of texts. To resolve this problem, we propose a spelling error correction model using a spelling error correction dictionary and a newspaper corpus. The proposed model has the advantage that the cost of data construction are not high because it uses a newspaper corpus, which we can easily obtain, as a training corpus. In addition, the proposed model has an advantage that additional external modules such as a morphological analyzer and a word-spacing error correction system are not required because it uses a simple string matching method based on a correction dictionary. In the experiments with a newspaper corpus and a short message corpus collected from real mobile phones, the proposed model has been shown good performances (a miss-correction rate of 7.3%, a F1-measure of 97.3%, and a false positive rate of 1.1%) in the various evaluation measures.

• Asia Pacific Journal of Corpus Research
• /
• v.2 no.1
• /
• pp.35-45
• /
• 2021

With the wide spread of the Common European Framework of Reference (CEFR) scales, many studies attempt to apply them in routine teaching and rater training, while more evidence regarding criterial features at different CEFR levels are still urgently needed. The current study aims to explore complexity features that distinguish and predict CEFR proficiency levels in oral performance. Using a quantitative/corpus-based approach, this research analyzed lexical and syntactic complexity features over 80 transcriptions (includes A1, A2, B1 CEFR levels, and native speakers), based on an interview test, Standard Speaking Test (SST). ANOVA and correlation analysis were conducted to exclude insignificant complexity indices before the discriminant analysis. In the result, distinctive differences in complexity between CEFR speaking levels were observed, and with a combination of six major complexity features as predictors, 78.8% of the oral transcriptions were classified into the appropriate CEFR proficiency levels. It further confirms the possibility of predicting CEFR level of L2 learners based on their objective linguistic features. This study can be helpful as an empirical reference in language pedagogy, especially for L2 learners' self-assessment and teachers' prediction of students' proficiency levels. Also, it offers implications for the validation of the rating criteria, and improvement of rating system.

• Journal of the Korean Society for information Management
• /
• v.35 no.2
• /
• pp.63-88
• /
• 2018

This research evaluated differences of classification performance for feature selection methods using LDA topic model and Doc2Vec which is based on word embedding using deep learning, feature corpus sizes and classification algorithms. In addition to find the feature corpus with high performance of classification, an experiment was conducted using feature corpus was composed differently according to the location of the document and by adjusting the size of the feature corpus. Conclusionally, in the experiments using deep learning evaluate training frequency and specifically considered information for context inference. This study constructed biomedical document dataset, Disease-35083 which consisted biomedical scholarly documents provided by PMC and categorized by the disease category. Throughout the study this research verifies which type and size of feature corpus produces the highest performance and, also suggests some feature corpus which carry an extensibility to specific feature by displaying efficiency during the training time. Additionally, this research compares the differences between deep learning and existing method and suggests an appropriate method by classification environment.