• Journal of Advanced Marine Engineering and Technology
• /
• v.34 no.6
• /
• pp.871-879
• /
• 2010

In this paper, we present a system that classifies comments on a news article into a user opinion called a polarity (positive or negative). The system is a kind of document classification system for comments and is based on machine learning techniques like support vector machine. Unlike normal documents, comments have their body that can influence classifying their opinions as polarities. In this paper, we propose a feature weighting scheme using such characteristics of comments and several resources for opinion classification. Through our experiments, the weighting scheme have turned out to be useful for opinion classification in comments on Korean news articles. Also Korean character n-grams (bigram or trigram) have been revealed to be helpful for opinion classification in comments including lots of Internet words or typos. In the future, we will apply this scheme to opinion analysis of comments of product reviews as well as news articles.

• Asian Journal of Innovation and Policy
• /
• v.8 no.3
• /
• pp.417-427
• /
• 2019

This article puts forward the opinion of a policy expert, who had been involved in Korea's science and technology policy for 30 years, about technology policy for SMEs. This article first explains why technology policies for SMEs are needed, and from what point of view. The next section looks at the current problems facing Korean SMEs, followed by the introduction of past and current polices to support SMEs. The comparison between current issues facing SMEs and past and current policies leads to lessons that can be learned. There are four lessons for the policy itself and three for the implementation of these policies. As for policy, the first four are about stages of development, concentration, R&D and diffusion, and policy mix. The latter three are policymakers, institutional building, and cooperation between ministries. This article makesfour suggestions for future policy: the importance of startups' culture, cooperation with others, infrastructure to reduce uncertainty, and policy monitoring and evaluation.

• Asian Journal for Public Opinion Research
• /
• v.1 no.3
• /
• pp.153-202
• /
• 2014

Aside from political leaders' popularity rates and the stock exchange index of business firms, ordinary people are highly interested in aspects of daily life, such as housing, income, health, family, food, human relations and work. Cross-national opinion polls on daily-life satisfaction were carried out in Japan, South Korea, Thailand, Hong Kong, Macao, the Philippines, India, Myanmar, Taiwan, China, Malaysia and Pakistan in the fall of 2013 and winter 2014. The percent difference index (PDI) is formulated as the sum of two positive responses (satisfied and somewhat satisfied) minus the sum of two negative responses (dissatisfied and somewhat dissatisfied). Percent difference indices are given according to society and daily-life aspects. For our analysis to go beneath national average and to go beyond national borders, two lines of analysis are carried out. First, the distance between the level of satisfaction of the top and bottom quartiles is given for each society and according to each of the daily-life aspects. Second, the regional sum of satisfaction of the top quartiles and bottom quartiles are shown crossed by daily-life aspects. In this article we confine ourselves to preliminary comparative description and analysis. More solid and deep comparisons will be carried out by local polling leaders of 12 Asian societies in the succeeding issue of the Asian Journal of Public Opinion Research. Nevertheless, two key threads stand out from this preliminary comparisons. First, social relations (family and human relations) stand out as most satisfied aspects of life in most of twelve societies. Second, the need to go beneath national averages and beyond national borders in analyzing cross-national surveys is confirmed. The comparability and validity of cross-national surveys with varying sampling method and survey mode are briefly discussed toward the end of the article.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.24 no.2
• /
• pp.335-343
• /
• 2014

It is more important to ensure the credibility of internet media which influence the public opinion. However, there are vague suspicions in public from the examples of manipulation of online reviews with anonymity. In this study, we explore the possibility of manipulating public opinion in online web sites. We investigate the characteristics of comments posted by users on web sites and compare each comments by using the cosine similarity function. Our result shows followings. First, we found a correlation between the similarities of comments and the article ranks in the web sites. Second, it is possible to identify abnormal user activities indicating excessive multiple posting, double posting and astroturf activities.

• The Journal of the Convergence on Culture Technology
• /
• v.7 no.2
• /
• pp.149-156
• /
• 2021

Text mining is one of the useful tools to discover public opinion and perception regarding political issues from big data. It is very common that users of social media express their opinion with newly-coined words such as slang and emoji. However, those new words are not effectively captured by traditional text mining methods that process text data using a language dictionary. In this study, we propose effective methods to extract newly-coined words that connote the political stance and opinion of users. With various text mining techniques, I attempt to discover the context and the political meaning of the new words.

• Asian Journal for Public Opinion Research
• /
• v.2 no.1
• /
• pp.3-7
• /
• 2014

How satisfied or dissatisfied people are with aspects of daily life such as housing, income, health, family, food, human relations, and work provides vital information about them. On the basis of a nation-wide random sample survey in Japan in October 2013, this article analyzes Japanese citizens' daily life satisfaction in a snapshot. The big picture of Japanese daily life is that they are more or less satisfied with their daily life, although income, work, and housing register lower levels of satisfaction in comparison to other aspects like food, family, human relations, and health. Their satisfaction is focused on food, family, and human relations.

• The Journal of Korean Medical History
• /
• v.14 no.1
• /
• pp.17-30
• /
• 2001

This research is an analysis of medical essay article 62 by Jang Jong-jung that is quotated in "Donguibogam". The opinion through out the essay is that the way of healing on the whole is eliminating the origin of illness above all. Three methods of sweating, vomiting, and purging that Jang applied are profoundly connected to this opinion. Non-medicinal treatment of emotion disorders that is quoted in "Donguibogam" is also closely related. The consideration of Unqi(circulating qi) as the best way of searching etiological factor, pathology, and treatment, in "Donguibogam" is close to the view of medical essay by Jang Jong-jung.

• Asian Journal for Public Opinion Research
• /
• v.2 no.3
• /
• pp.196-227
• /
• 2015

The premise of the intentional model of split-ticket voting is that some voters split their tickets simply because they prefer divided government and believe in constant "checks and balances." This article examines whether this premise stands firm in an emerging democracy like Taiwan. That is, by using survey data in Taiwan, we explore whether one's attitude toward divided or unified government is "real." We hypothesize that a citizen's attitude toward "checks and balances" is subject to change, and conditional on whether her preferred party is in power. Specifically, we speculate that a citizen would tend to hold the balancing perspective or favor divided government, if her preferred party is in opposition. However, if her preferred party becomes the ruling party, she would be more likely to oppose (hold) the balancing (non-balancing) perspective or favor unified government. We then utilize panel survey data embedded in Taiwan's Election and Democratization Studies (TEDS) to verify our hypothesis.

• Tribology and Lubricants
• /
• v.13 no.2
• /
• pp.8-15
• /
• 1997

It is very difficult to express the state of a tested engine objectively concerning the malfunction, failure and wear of an engine. The general method of engine evaluation is to express evaluator's opinion for the engine state after testing. However, these methods is only subjective evaluation because test engineers, designers, and evaluators does not easily coincide in opinion. Therefore, in this article, the method of engine rating are introduced in order to represent the qualitative trends into quantitative value. The purpose of Engine Rating is to assign a quantified value to the tribology state of a tested engines. Originally, this Engine Rating method have been used to evaluate the performance of engine oil quantitavely. Using this method, we can predict the proper interval of engine oil change due to its objectivity. So, we can prevent the frequent change of engine oil and protect the environmental contamination. Furthermore, this method can be used to tell the general state of a tested engine after finishing engine durability tests. A single merit scale is used to evaluate numerically the state of cleanliness and the mechanical condition. Generally, a part which is absolutely clean or shows no wear, is rated merit 10. A part which is absolutely dirty or very worn, is rated merit 0.

• Journal of the Korea Society of Computer and Information
• /
• v.23 no.1
• /
• pp.9-15
• /
• 2018

Policy polls, which investigate the degree of support that the policy has for policy implementation, play an important role in making decisions. As the number of Internet users increases, the public is actively commenting on their policy news stories. Current policy polls tend to rely heavily on phone and offline surveys. Collecting and analyzing policy articles is useful in policy surveys. In this study, we propose a method of analyzing comments using deep learning technology showing outstanding performance in various fields. In particular, we designed various models based on the recurrent neural network (RNN) which is suitable for sequential data and compared the performance with the support vector machine (SVM), which is a traditional machine learning model. For all test sets, the SVM model show an accuracy of 0.73 and the RNN model have an accuracy of 0.83.