• Journal of KIISE:Software and Applications
• /
• v.29 no.4
• /
• pp.270-277
• /
• 2002

Text filtering is a task of deciding whether a document has relevance to a specified topic. As Internet and Web becomes wide-spread and the number of documents delivered by e-mail explosively grows the importance of text filtering increases as well. The aim of this paper is to improve the accuracy of text filtering systems by using machine learning techniques. We apply AdaBoost algorithms to the filtering task. An AdaBoost algorithm generates and combines a series of simple hypotheses. Each of the hypotheses decides the relevance of a document to a topic on the basis of whether or not the document includes a certain word. We begin with an existing AdaBoost algorithm which uses weak hypotheses with their output of 1 or -1. Then we extend the algorithm to use weak hypotheses with real-valued outputs which was proposed recently to improve error reduction rates and final filtering performance. Next, we attempt to achieve further improvement in the AdaBoost's performance by first setting weights randomly according to the continuous Poisson distribution, executing AdaBoost, repeating these steps several times, and then combining all the hypotheses learned. This has the effect of mitigating the ovefitting problem which may occur when learning from a small number of data. Experiments have been performed on the real document collections used in TREC-8, a well-established text retrieval contest. This dataset includes Financial Times articles from 1992 to 1994. The experimental results show that AdaBoost with real-valued hypotheses outperforms AdaBoost with binary-valued hypotheses, and that AdaBoost iterated with random weights further improves filtering accuracy. Comparison results of all the participants of the TREC-8 filtering task are also provided.

• Science of Emotion and Sensibility
• /
• v.8 no.3
• /
• pp.221-229
• /
• 2005

This study aimed to understand the sex differences in cognitive behaviors in website design and demonstrate a practical basis for utilizing these differences into more user-centered design concept. Especially, we focused on the sex-different preference according to the information architecture of website navigation. First, We investigated general differences between men and women in cognitive behaviors through various literature studies. According to our investigation, men's cognitive works generally tend to follow a regular sequence and proceed step by step. On the other hand, women's cognitive style is generally characterized by random generation and simultaneous progress. To examine that these differences can be found in use of website navigation, we made an experiment in website design. We designed several test websites that have same contents but different style of navigation structure. A similar number of men and women were chosen for this test and they implemented given tasks. During the test, participants reported their preference on each websites and their implementing time and number of errors were collected. Based on the analysis of test data, it was possible to conclude that male participants' preference for the navigation with a narrow and deep information structure is relatively higher than female participants' preference for the same navigation, On the other hand, female participants have a preference of the navigation with a broad and swallow information structure. The result of study showed that there is a close correlation between the sex differences in preference of navigation types and the general sex differences in cognitive behavior. This finding can be used as a basis for designing the website navigation in which sex differences are reflected.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.40 no.7
• /
• pp.1296-1306
• /
• 2015

Nowadays the research of energy saving on the IP networks have been studied the various methods in many research institutes. This paper suggests the energy saving method in IP networks which have the various energy profiles, and analyzes its energy saving characteristics in detail. Especially this paper proposes the energy profile based OSPF routing method which have the selectable weighted value in OSPF metric and energy consumption in IP network. This paper analyzes the energy saving effects of the various situations to minimize the energy consumption using the various weighted value on the proposed scheme. The results show that the energy saving efficiency can get about 67% at in ingress input load ${\rho}=0.5$ by using random energy profiles in IP networks. Although the number of hops is a slight increased due to routing the paths for the minimum energy consumption in the algorithm of this method, the increment hop number is limited the mean 1.4 hops. This paper confirms that the energy profile of core router has the large effects of energy saving than the energy profile of edge router, and the proposed method has the excellent energy saving characteristics in IP networks.

• Journal of Communications and Networks
• /
• v.16 no.2
• /
• pp.227-237
• /
• 2014

Throughput scaling laws for two coexisting ad hoc networks with m primary users (PUs) and n secondary users (SUs) randomly distributed in an unit area have been widely studied. Early work showed that the secondary network performs as well as stand-alone networks, namely, the per-node throughput of the secondary networks is ${\Theta}(1/\sqrt{n{\log}n})$. In this paper, we show that by exploiting directional spectrum opportunities in secondary network, the throughput of secondary network can be improved. If the beamwidth of secondary transmitter (TX)'s main lobe is ${\delta}=o(1/{\log}n)$, SUs can achieve a per-node throughput of ${\Theta}(1/\sqrt{n{\log}n})$ for directional transmission and omni reception (DTOR), which is ${\Theta}({\log}n)$ times higher than the throughput with-out directional transmission. On the contrary, if ${\delta}={\omega}(1/{\log}n)$, the throughput gain of SUs is $2{\pi}/{\delta}$ for DTOR compared with the throughput without directional antennas. Similarly, we have derived the throughput for other cases of directional transmission. The connectivity is another critical metric to evaluate the performance of random ad hoc networks. The relation between the number of SUs n and the number of PUs m is assumed to be $n=m^{\beta}$. We show that with the HDP-VDP routing scheme, which is widely employed in the analysis of throughput scaling laws of ad hoc networks, the connectivity of a single SU can be guaranteed when ${\beta}$ > 1, and the connectivity of a single secondary path can be guaranteed when ${\beta}$ > 2. While circumventing routing can improve the connectivity of cognitive radio ad hoc network, we verify that the connectivity of a single SU as well as a single secondary path can be guaranteed when ${\beta}$ > 1. Thus, to achieve the connectivity of secondary networks, the density of SUs should be (asymptotically) bigger than that of PUs.

• The Journal of the Korea Contents Association
• /
• v.20 no.1
• /
• pp.356-363
• /
• 2020

As a becoming era of Internet-of-Things, various devices are connected via wire or wirless networks. Although every day life is more convenient, security problems are also increasing such as privacy, information leak, denial of services. Since ECC, a kind of public key cryptosystem, has a smaller key size compared to RSA, it is widely used for environmentally constrained devices. The key of ECC in constrained devices can be exposed to power analysis attacks during scalar multiplication operation. In this paper, a key-randomization method is suggested for scalar multiplication on SECG parameters. It is against differential power analysis and has operational efficiency. In order to increase of operational efficiency, the proposed method uses the property 2^lP=∓cP where the constant c is small compared to the order n of SECG parameters and n=2^l±c. The number of operation for the Coron's key-randomization scalar multiplication algorithm is 21, but the number of operation for the proposed method in this paper is (3/2)l. It has efficiency about 25% compared to the Coron's method using full random numbers.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.62 no.4
• /
• pp.33-43
• /
• 2020

In order to reduce damage from farmland inundation caused by recent climate change, it is necessary to predict the risk of farmland inundation accurately. Inundation modeling should be performed by considering multiple time distributions of possible rainfalls, as digital forecasts of Korea Meteorological Administration is provided on a six-hour basis. As building multiple inputs and creating inundation models take a lot of time, it is necessary to shorten the forecast time by building a data base (DB) of farmland inundation probability. Therefore, the objective of this study is to establish a DB of farmland inundation probability in accordance with forecasted rainfalls. In this study, historical data of the digital forecasts was collected and used for time division. Inundation modeling was performed 100 times for each rainfall event. Time disaggregation of forecasted rainfall was performed by applying the Multiplicative Random Cascade (MRC) model, which uses consistency of fractal characteristics to six-hour rainfall data. To analyze the inundation of farmland, the river level was simulated using the Hydrologic Engineering Center - River Analysis System (HEC-RAS). The level of farmland was calculated by applying a simulation technique based on the water balance equation. The inundation probability was calculated by extracting the number of inundation occurrences out of the total number of simulations, and the results were stored in the DB of farmland inundation probability. The results of this study can be used to quickly predict the risk of farmland inundation, and to prepare measures to reduce damage from inundation.

• Journal of Korea Technology Innovation Society
• /
• v.21 no.2
• /
• pp.522-559
• /
• 2018

In this study, we reviewed and analyzed the influencing factors of technology transfer performance in the previous studies (52 domestic journals and theses) and classified the various influencing factors into 6 top factors and 13 sub-factors based on the theoretical background. The study results of previous articles were analyzed by meta-analysis method so as to calculate the overall average effect size of influencing factors of technology transfer performance. As the result, the overall effect size (ESr) calculated through meta-analysis applying random effect model is .269, which corresponds to the medium effect size. By comparing effect sizes of influencing factors, the four(4) key influencing factors were also identified, which are 'number of researchers', 'dedicated organization', 'possess technology', and 'external cooperation'. The technology transfer performance are divided into three types: the number of technology transfers, technology transfer income, and other technology transfer performances. The major influencing factors of each type are derived through meta-analysis at the sub-category level. As moderator variables, the paper type and the data type were analyzed but no significant results were obtained. Since this research is limited to the technology transfer, it is necessary to carry out the study related to the influencing factors on the technology commercialization as following study.

• Tuberculosis and Respiratory Diseases
• /
• v.42 no.4
• /
• pp.513-521
• /
• 1995

Background: Nucleolar organizer regions(NORs) are chromosomal segments encoding for ribosomal RNA and associated with argyrophilic nonhistone protein. Ribosomal RNA genes ultimately direct ribosome and protein synthesis, and it has been suggested the numbers of NORs detected in the cell may reflect nuclear and cellular activity. This study was performed to evaluate the applicability of AgNORs to the diagnosis of squamous cell carcinoma of the lung. Method: The one step silver methods(AgNORs) was used to stain NORs in the routinely processed, formalin fixed, paraffin embedded sections of 36 cases of squamous cell carcinoma of the lung obtained by surgical resection of primary tumor. In each specimen, 100 tumor cells and 100 normal cells adjacent to the tumor chosen at random were examined under an oil immersion lens at a magnification of ${\times}1000$. The mean number of AgNORs per nucleus was calculated for each specimen. Results: The mean number of AgNORs per nucleus(mAgNORs) of normal bronchial epithelium and squamous cell carcinoma of the lung was $1.74{\pm}0.25$ and $4.05{\pm}0.80$, respectively. The difference of mAgNOR between normal and tumor tissue was statistically significant(p<0.001). There was no statistical difference among tumors of different stages. The difference of mAgNOR between normal and tumor tissue was statistically significant in each TNM stage(p<0.05). Conclusion: Mean AgNOR count may be used as a useful marker for the differential diagnosis of benignancy and malignancy, and proliferative activity of the cell in squamous cell carcinoma of the lung. But there was no statistical difference in mean AgNOR count among tumors of different surgical stages. Further studies for the application of mAgNORs to the diagnosis of other histologic types and cytologic specimens of the lung cancer are needed.

• Journal of Preventive Medicine and Public Health
• /
• v.27 no.3 s.47
• /
• pp.475-486
• /
• 1994

In order to provide necessary information for better health of workers through understanding the actual status of the industries adopting shift systems. The data were gathered from a stratified random sample of 347 (0.5%) firms selected out of about 70,000 manufacturing industries throughout the nation in 1993. Stratification during sampling was by industrial group and number of workers. The major findings obtained from 288 firms surveyed completely were as follows : 1. About 20.2% of the 288 firms were adopting shift systems and shirtworkers accounted for about 25.1% of the total work force in 288 firms. 2. The bigger number of workers was, the higher the adopting rate of shift system was. 3. The rate of having welfare facilities such as dining room, commuting bus, washing facilities, and health care room etc. was higher in the industries adopting shift systems. 4. The major industrial groups adopting shift systems were the rubber a: plastic goods manufacturing industry (54.1 per 100 firms) and the textile manufacturing industry (44.8 per 100 firms). However, the proportion of shiftworkers was higher in the textile manufacturing industry (70.5 per 100 firms) and the electronics industry (57.9 per 100 frms). 5. The most predominant work schedule was the weekly rotating, semi-continuos 2-crew 2-shift system (47.5%). 6. In the industries adopting shift systems, shiftworkers had an adjusted average of 0.29 spells per 100 workers ($0.14{\sim}0.45$ in 95% C.I.) compared to 0.23 spells per 100 nonshift dayworkers ($0.15{\sim}0.31$ in 95% C.I.) for 1 month. 7. Also, in the industries adopting shift systems, the adjusted average annuel turn-over rate of shiftworkers was 13.07 per 100 workers ($12.03{\sim}14.12$ in 95% C.I.) compared to 10.18 per 100 nonshift dayworkers ($9.53{\sim}10.82$ in 95% C.I.).

• Journal of the HCI Society of Korea
• /
• v.5 no.1
• /
• pp.37-43
• /
• 2010

Twitter has been continuously growing since October, 2006. Especially, not only the users and the number of messages have been increasing but also a new concept in social networking called 'micro blogging' has diffused. Within Korea, service such as 'me2day' has already been introduced and the improvement of internet accessibility within mobile devices is expected to expand the 'micro blogs'. In this point, this research is executed to study the new medium, 'micro blog'. To do so, we collected and analyzed Twitter logs of Korean users. Especially, we were curious about the extreme heavy users using Twitter, despite of the linguistic and cultural barrier of the foreign service. Who they are, why and how they use the 'micro blog'. First, we reviewed the general aspect of followers and messages by collecting a certain number of random samples. Using the Lorenz curve we found out that there was the imbalance within the users and based on this phenomenon we deducted an extreme heavy user group. In order to perform further analysis, log analysis was performed on the extreme heavy users. As the result, the users used multiple mobile and desktop 'Twitter' clients. The usage pattern was similar to that of internet usage time but was used during their "micro" time. The users using 'Twitter' not only to spread messages about important information, special events and emotions, but also as a habitual 'chatting tool' to express ordinary personal chats similar to SMS and IM services. In this research, it is proved that 68% of the total messages were ordinary personal chats. Also, with 24% of the total messages were retweets, we were able to find out that virtually connected 'people' and 'relationships' acted as the dominant trigger of their articulation.