• Journal of KIISE:Software and Applications
• /
• v.29 no.9
• /
• pp.654-668
• /
• 2002

In this paper, we present an empirical study for improving the Korean text chunking based on machine learning and feature set selection approaches. We focus on two issues: the problem of selecting feature set for Korean chunking, and the problem of alleviating the data sparseness. To select a proper feature set, we use a heuristic method of searching through the space of feature sets using the estimated performance from a machine learning algorithm as a measure of "incremental usefulness" of a particular feature set. Besides, for smoothing the data sparseness, we suggest a method of using a general part-of-speech tag set and selective lexical information under the consideration of Korean language characteristics. Experimental results showed that chunk tags and lexical information within a given context window are important features and spacing unit information is less important than others, which are independent on the machine teaming techniques. Furthermore, using the selective lexical information gives not only a smoothing effect but also the reduction of the feature space than using all of lexical information. Korean text chunking based on the memory-based learning and the decision tree learning with the selected feature space showed the performance of precision/recall of 90.99%/92.52%, and 93.39%/93.41% respectively.

• Proceedings of the Korean Information Science Society Conference
• /
• 2000.10b
• /
• pp.197-199
• /
• 2000

연속음성인식을 수행하기 위해서는 발음사전과 언어모델이 필요하다. 이 둘 사이에는 디코딩 단위가 일치하여야 하므로 발음사전 구축시 디코딩 단위로 표제어 단위를 선정하며 표제어 사이의 음운변화 현상을 반영한 발음사전을 구축하여야 한다. 한국어에 부합하는 음운변화현상을 분석하여 학습용 자동 발음열을 생성하고, 이를 통하여 발음사전을 구축한다. 전처리 단계로 기호, 단위, 숫자 등 전처리 과정 및 형태소 분석 과정을 수행하며, 디코딩 단위인 의사 형태소 단위를 생성하기 위해 규칙을 이용한 태깅 과정을 거친다. 이를 통해 나온 결과를 발음열 생성기 입력으로 하며, 결과는 학습용 발음열 또는 발음사전 구성을 위한 형태로 출력한다. 표제어간 음운변화 현상이 반영된 상태의 표제어 단위이므로 실제 음운변화가 반영되지 않은 상태의 표제어와는 그 형태가 상이하다. 이는 연속 발음시 생기는 현상으로 실제 인식에는 이 음운변화 현상이 반영된 사전이 필요하게 된다. 생성된 발음사전의 효용성을 확인하기 위해 다음과 같은 실험을 통해 성능을 평가하였다. 음향학습을 위하여 PBS(Phonetically Balanced Sentence) 낭독체 17200문장을 녹음하고 그 전사파일을 사용하여 학습을 수행하였고, 발음사전의 평가를 위하여 이 중 각각 3100문장을 사용하여 다음과 같은 실험을 수행하였다. 형태소 태그정보를 이용하여 표제어간 음운변화 현상을 반영한 최적의 발음사전과 다중 발음사전, 언어학적 기준에 의한 수작업으로 생성한 표준 발음사전, 그리고 표제어간의 음운변화 현상을 고려하지 않고 독립된 단어로 생성한 발음사전과의 비교 실험을 수행하였다. 실험결과 표제어간 음운변화 현상을 반영하지 않은 경우 단어 인식률이 43.21%인 반면 표제어간 음운변화 현상을 반영한 1-Best 사전의 경우 48.99%, Multi 사전의 경우 50.19%로 인식률이 5~6%정도 향상되었음을 볼 수 있었고, 수작업에 의한 표준발음사전의 단어 인식률 45.90% 보다도 약 3~4% 좋은 성능을 보였다.

• Journal of The Korean Society of Emergency Medicine
• /
• v.29 no.5
• /
• pp.437-448
• /
• 2018

Objective: Acute ischemic stroke (AIS) requires time-dependent reperfusion therapy, and early recognition of AIS is important to patient outcomes. This study was conducted to identify the clinical features and risk factors of AIS patients that are missed during the early stages of diagnosis. Methods: We retrospectively reviewed AIS patients admitted to a hospital through the emergency department. AIS patients were defined as ischemic stroke patients who visited the emergency department within 6 hours of symptom onset. Patients were classified into two groups: an activation group (A group), in which patients were identified as AIS and the stroke team was activated, and a non-activation group (NA group), for whom the stroke team was not activated. Results: The stroke team was activated for 213 of a total of 262 AIS patients (81.3%), while it was not activated for the remaining 49 (18.7%). The NA group was found to be younger, have lower initial National Institutes of Health Stroke Scale scores, lower incidence of previous hypertension, and a greater incidence of cerebellum and cardio-embolic infarcts than the A group. The chief complaints in the A group were traditional stroke symptoms, side weakness (61.0%), and speech disturbance (17.8%), whereas the NA group had non-traditional symptoms, dizziness (32.7%), and decreased levels of consciousness (22.4%). Independent factors associated with missed stroke team activation were nystagmus, nausea/vomiting, dizziness, gait disturbance, and general weakness. Conclusion: A high index of AIS suspicion is required to identify such patients with these findings. Education on focused neurological examinations and the development of clinical decision tools that could differentiate non-stroke and stroke are needed.

• Journal of the Architectural Institute of Korea Planning & Design
• /
• v.34 no.11
• /
• pp.75-84
• /
• 2018

This study analyzed users' acceptance and intention to use in addition to needs and preferences of smart home technologies, and identified the differences in technology preference and acceptance by different factors. The subjects were residents in the 40s and 60s residing in the Seoul or suburbs of Seoul, and questionnaires were conducted in the 40s while interviews with questionnaires were conducted in the 60s. A total of 105 questionnaires were used as data, and frequency, mean, crossover, independent sample t test, one-way ANOVA and multiple regression analysis were performaed using SPSS23. The results of this study are as follows. First, hypertension, hyperlipidemia and hypercholesterolemia were the most common diseases among respondents and if there was no discomfort, they would like to continue living in the homes of the current residence. Therefore, the direction of smart home development should support the daily living and health care so that residents can live a healthy life for a long time in their living space. Second, the technologies that residents most need were a control technology of residential environments and a monitoring technology of residents' health and physiological changes. The most preferred sensor types are motion sensors and speech recognition while video cameras have a very low preference. Third, technology anxiety was the most significant factor influencing intention to accept smart home technology. The greater the technology anxiety is, the weaker the acceptance of technology. Fourth, when applying smart residential technology in homes, various resident characteristics should be considered. Age and technology intimacy were the most influential variables, and accordingly there were differences in technology preference and acceptance. Therefore, a user-friendly smart home plan should be done in the consideration of the results.

• Journal of Broadcast Engineering
• /
• v.25 no.6
• /
• pp.854-860
• /
• 2020

Deep neural networks have shown remarkable performance in various areas, including image classification and speech recognition. The variety of data generated by augmentation plays an important role in improving the performance of the neural network. The transformation of data in the augmentation process makes it possible for neural networks to be learned more generally through more diverse forms. In the traditional field of image process, not only new augmentation methods have been proposed for improving the performance, but also exploring methods for an optimal augmentation policy that can be changed according to the dataset and structure of networks. Inspired by the prior work, this paper aims to explore to search for an optimal augmentation policy in the field of sound data. We carried out many experiments randomly combining various augmentation methods such as adding noise, pitch shift, or time stretch to empirically search which combination is most effective. As a result, by applying the optimal data augmentation policy we achieve the improved classification accuracy on the environmental sound classification dataset (ESC-50).

• The Journal of the Convergence on Culture Technology
• /
• v.7 no.1
• /
• pp.234-239
• /
• 2021

The purpose of this paper is to find out the nature of language variation by exploring the ways of the progress of the language variation that produces all English-lects, i.e., the World Englishes. The study of language variation in linguistics is a hybrid enterprise, so the study of World Englishes has led to the recognition of a highly diverse set of all English-lects, encompassing regional dialects, sociolects, ethnolects and (post-)colonial dialects of World Englishes. In this paper, we propose a hybrid language variation model with three interacting factors of social distancing, on/off-contact, and linguistic diversity to examine the characteristics of language variation. In the context of World Englishes, the social distance is typically low in terms of their local location (country/speech) for local purposes. The social distance also varies based on online/offline communication modes and other social factors like gender, age and ethnic groups, resulting in all English-lects. To clarify the nature of World Englishes, the core Englishes, BrE, AmE and CanE are discussed here.

• International Journal of Advanced Culture Technology
• /
• v.10 no.1
• /
• pp.24-35
• /
• 2022

The purpose of this study was to form a linked model in which local institutions related to lifelong education for the disabled can cooperate based on the Daegu University K-PACE Center. The contents of the study started with recognizing the problem that the adult-centered lifelong education support system does not effectively cope with these factors, even though the independent life of people with developmental disabilities is a major factor determining the quality of life. Regarding this problem recognition, this study primarily emphasized the view that educational support for independent life of people with developmental disabilities should establish the context of the school foundation. The context of the school foundation is established for lifelong education centered on adulthood for people with developmental disabilities because the curriculum is embodied through the standards of subject matter education. In this regard, the Daegu University K-PACE Center, which established a curriculum that supports the independent life of people with developmental disabilities in terms of linking higher and lifelong education, actually reflects the context of the school foundation. As a result, this study prepared a strategy that could be considered as a transition to advance the curriculum organized by the Daegu University K-PACE Center, and the strategy was secondarily reflected as a procedure that could be linked to local lifelong education-related institutions for the disabled. Finally, this study presented a form of transition in which people with developmental disabilities can access the curriculum of lifelong education through the connection of local lifelong education-related institutions for the disabled, centering on the entire life of adulthood.

• Smart Media Journal
• /
• v.10 no.4
• /
• pp.71-79
• /
• 2021

One of the social aspects that changes as the use of the Internet becomes widespread is communication in online space. In the past, only one-on-one conversations were possible remotely, except when they were physically in the same space, but nowadays, technology has been developed to enable communication with a large number of people remotely through bulletin boards, communities, and social network services. Due to the development of such information and communication networks, life becomes more convenient, and at the same time, the damage caused by rapid information exchange is also constantly increasing. Recently, cyber crimes such as sending sexual messages or personal attacks to certain people with recognition on the Internet, such as not only entertainers but also influencers, have occurred, and some of those exposed to these cybercrime have committed suicide. In this paper, in order to reduce the damage caused by malicious comments, research a method for improving the performance of discriminate malicious comments through feature extraction based on parts-of-speech.

• International Journal of Computer Science & Network Security
• /
• v.22 no.6
• /
• pp.7-12
• /
• 2022

The modern stage of society's development is characterized by the rapid penetration of information technologies into all spheres of life. Their use contributes to improving the quality of tourism services, as well as the competitiveness of tourism industry enterprises. The role of information technology in tourism is growing more and more every year, which determines the relevance of the study of modern trends in the use of information technology in the tourism sector. The purpose of the study is to determine the possibilities of using information technologies to improve the quality of services provided to the tourism industry under the COVID-19 restrictions. The article systematizes the main approaches to the "cluster" category and provides an original definition of the "regional tourist cluster" concept. Based on an expert survey, the main trends in the introduction of information technologies in the tourism industry under the COVID-19 restrictions have been identified, which include virtual reality and augmented reality, speech recognition technologies, photo, video, audio (contactless control technologies), mobile IT applications and Big Data technologies. It has been concluded that the vast majority of improvements in the organization of tourism services under restrictions will be based on the organization of virtual solutions and online activities. The types of tourism services will also change, and information technology will help their development and dissemination.

• Asia-pacific Journal of Multimedia Services Convergent with Art, Humanities, and Sociology
• /
• v.7 no.10
• /
• pp.935-943
• /
• 2017

CNN (Convolution neural network), which is used for image classification and speech recognition among neural networks learning based on positive data, has been continuously developed to have a high performance structure to date. There are many difficulties to utilize in an embedded system with limited resources. Therefore, we use GPU (General-Purpose Computing on Graphics Processing Units), which is used for general-purpose operation of GPU to solve the problem because we use pre-learned weights but there are still limitations. Since CNN performs simple and iterative operations, the computation speed varies greatly depending on the thread allocation and utilization method in the Single Instruction Multiple Thread (SIMT) based GPGPU. To solve this problem, there is a thread that needs to be relaxed when performing Convolution and Pooling operations with threads. The remaining threads have increased the operation speed by using the method used in the following feature maps and kernel calculations.