• 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.

• Journal of the Korean Data and Information Science Society
• /
• v.28 no.2
• /
• pp.297-307
• /
• 2017

The epidemic model is used to model the spread of disease and to control the disease. In this research, we utilize SEIR model which is one of applications the SIR model that incorporates Exposed step to the model. The SEIR model assumes that a people in the susceptible contacted infected moves to the exposed period. After staying in the period, the infectee tends to sequentially proceed to the status of infected, recovered, and removed. This type of infection can be used for research in cases where there is a latency period after infectious disease. In this research, we collected respiratory infectious disease data for the Middle East Respiratory Syndrome Coronavirus (MERSCoV). Assuming that the spread of disease follows a stochastic process rather than a deterministic one, we utilized the Poisson process for the variation of infection and applied epidemic model to the stochastic chemical reaction model. Using observed pandemic data, we estimated three parameters in the SIER model; exposed rate, transmission rate, and recovery rate. After estimating the model, we applied the fitted model to the explanation of spread disease. Additionally, we include a process for generating the Exposed trajectory during the model estimation process due to the lack of the information of exact trajectory of Exposed.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.17 no.8
• /
• pp.1947-1954
• /
• 2013

Previous works for detecting arrhythmia have mostly used nonlinear method such as artificial neural network, fuzzy theory, support vector machine to increase classification accuracy. Most methods require accurate detection of P-QRS-T point, higher computational cost and larger processing time. But it is difficult to detect the P and T wave signal because of person's individual difference. Therefore it is necessary to design efficient algorithm that classifies different arrhythmia in realtime and decreases computational cost by extrating minimal feature. In this paper, we propose arrhythmia detection based on binary coding using QRS feature varibility. For this purpose, we detected R wave, RR interval, QRS width from noise-free ECG signal through the preprocessing method. Also, we classified arrhythmia in realtime by converting threshold variability of feature to binary code. PVC, PAC, Normal, BBB, Paced beat classification is evaluated by using 39 record of MIT-BIH arrhythmia database. The achieved scores indicate the average of 97.18%, 94.14%, 99.83%, 92.77%, 97.48% in PVC, PAC, Normal, BBB, Paced beat classification.

• Research in Plant Disease
• /
• v.26 no.4
• /
• pp.239-249
• /
• 2020

To find out the cause of the fire blight outbreak in apples and pears of Korea in 2019, we investigated disease appearing situation of thirty fire blight infected orchards, and interviewed farmers to determine the cultivation characteristics. Fire blight occurred mostly in orchards that had infected more than 2 years before. The cause of this were as follows: farmers did not know the symptoms of the disease properly. It is presumed that it has spread from the first occurrence to the surrounding orchards by flower-visiting insects or farmers and to a short distance or a long distance by the same cultivator or co-farmer. These series of processes repeated in the newly spreading area, and then disease reports increased as farmers became aware of fire blight. To minimize the spread of fire blight in Korea, it suggested that thorough education of farmers for early diagnosis and quantitative detection technology that can diagnose even in no symptom showing plants. And chemical or biological spraying systems suitable for domestic cultivation methods, which are producing large fruits, and molecular epidemiological studies of pathogens.

• Journal of Internet Computing and Services
• /
• v.17 no.1
• /
• pp.23-33
• /
• 2016

The current smart lighting is shaped to offer the lighting environment suitable for current context, after identifying user's action and location through a sensor. The sensor-based context awareness technology just considers a single user, and the studies to interpret many users' various context occurrences and conflicts lack. In existing studies, a fuzzy theory and algorithm including ReBa have been used as the methodology to solve context conflict. The fuzzy theory and algorithm including ReBa just avoid an opportunity of context conflict that may occur by providing services by each area, after the spaces where users are located are classified into many areas. Therefore, they actually cannot be regarded as customized service type that can offer personal preference-based context conflict. This paper proposes a priority-based LED lighting control system interpreting multiple context conflicts, which decides services, based on the granted priority according to context type, when service conflict is faced with, due to simultaneous occurrence of various contexts to many users. This study classifies the residential environment into such five areas as living room, 'bed room, study room, kitchen and bath room, and the contexts that may occur within each area are defined as 20 contexts such as exercising, doing makeup, reading, dining and entering, targeting several users. The proposed system defines various contexts of users using an ontology-based model and gives service of user oriented lighting environment through rule based on standard and context reasoning engine. To solve the issue of various context conflicts among users in the same space and at the same time point, the context in which user concentration is required is set in the highest priority. Also, visual comfort is offered as the best alternative priority in the case of the same priority. In this manner, they are utilized as the criteria for service selection upon conflict occurrence.