• Proceedings of the Korean Information Science Society Conference
• /
• 2006.10c
• /
• pp.222-225
• /
• 2006

한국은 다른 나라에 비해 많은 인터넷 사용자를 가지고 있다. 이에 비례해서 한국의 인터넷 유저들은 Spam Mail에 대해 많은 불편함을 호소하고 있다. 이러한 문제를 해결하기 위해 본 논문은 다양한 Feature Weighting, Feature Selection 그리고 문서 분류 알고리즘들을 이용한 한국어 스팸 문서 Filtering연구에 대해 기술한다. 그리고 한국어 문서(Spam/Non-Spam 문서)로부터 영사를 추출하고 이를 각 분류 알고리즘의 Input Feature로써 이용한다. 그리고 우리는 Feature weighting 에 대해 기존의 전통적인 방법이 아니라 각 Feature에 대해 Variance 값을 구하고 Global Feature를 선택하기 위해 Max Value Selection 방법에 적용 후에 전통적인 Feature Selection 방법인 MI, IG, CHI 들을 적용하여 Feature들을 추출한다. 이렇게 추출된 Feature들을 Naive Bayes, Support Vector Machine과 같은 분류 알고리즘에 적용한다. Vector Space Model의 경우에는 전통적인 방법 그대로 사용한다. 그 결과 우리는 Support Vector Machine Classifier, TF-IDF Variance Weighting(Combined Max Value Selection), CHI Feature Selection 방법을 사용할 경우 Recall(99.4%), Precision(97.4%), F-Measure(98.39%)의 성능을 보였다.

• Journal of the Korean Society for Library and Information Science
• /
• v.39 no.1
• /
• pp.45-58
• /
• 2005

This Paper is a comparative study of feature selection methods for Korean web documents clustering. First, we focused on how the term feature and the co-link of web documents affect clustering performance. We clustered web documents by native term feature, co-link and both, and compared the output results with the originally allocated category. And we selected term features for each category using $X^2$, Information Gain (IG), and Mutual Information (MI) from training documents, and applied these features to other experimental documents. In addition we suggested a new method named Max Feature Selection, which selects terms that have the maximum count for a category in each experimental document, and applied $X^2$ (or MI or IG) values to each term instead of term frequency of documents, and clustered them. In the results, $X^2$ shows a better performance than IG or MI, but the difference appears to be slight. But when we applied the Max Feature Selection Method, the clustering Performance improved notably. Max Feature Selection is a simple but effective means of feature space reduction and shows powerful performance for Korean web document clustering.

• Journal of KIISE:Software and Applications
• /
• v.33 no.8
• /
• pp.692-702
• /
• 2006

In this study, a weighted fuzzy min-max (WFMM) neural network model for pattern classification is proposed. The model has a modified structure of FMM neural network in which the weight concept is added to represent the frequency factor of feature values in a learning data set. First we present in this paper a new activation function of the network which is defined as a hyperbox membership function. Then we introduce a new learning algorithm for the model that consists of three kinds of processes: hyperbox creation/expansion, hyperbox overlap test, and hyperbox contraction. A weight adaptation rule considering the frequency factors is defined for the learning process. Finally we describe a feature analysis technique using the proposed model. Four kinds of relevance factors among feature values, feature types, hyperboxes and patterns classes are proposed to analyze relative importance of each feature in a given problem. Two types of practical applications, Fisher's Iris data and Cleveland medical data, have been used for the experiments. Through the experimental results, the effectiveness of the proposed method is discussed.

• ETRI Journal
• /
• v.34 no.5
• /
• pp.743-752
• /
• 2012

Semantic interpretation of the relationship between noun compound (NC) elements has been a challenging issue due to the lack of contextual information, the unbounded number of combinations, and the absence of a universally accepted system for the categorization. The current models require a huge corpus of data to extract contextual information, which limits their usage in many situations. In this paper, a new semantic relations interpreter for NCs based on novel lightweight binary features is proposed. Some of the binary features used are novel. In addition, the interpreter uses a new feature selection method. By developing these new features and techniques, the proposed method removes the need for any huge corpuses. Implementing this method using a modular and plugin-based framework, and by training it using the largest and the most current fine-grained data set, shows that the accuracy is better than that of previously reported upon methods that utilize large corpuses. This improvement in accuracy and the provision of superior efficiency is achieved not only by improving the old features with such techniques as semantic scattering and sense collocation, but also by using various novel features and classifier max entropy. That the accuracy of the max entropy classifier is higher compared to that of other classifiers, such as a support vector machine, a Na$\ddot{i}$ve Bayes, and a decision tree, is also shown.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2007.11a
• /
• pp.167-171
• /
• 2007

본 논문은 패턴 분류 문제에 사용되는 퍼지 최대-최소 신경망 방법을 이용하여 특정 집합으로부터 새로운 특정 집합을 추출해내고 추출된 특정 집합으로부터 의미 있는 특정을 선택해 내는 새로운 방법을 제안한다. 퍼지 최대-최소 신경망은 패턴 분류를 위해 주로 사용이 되어 왔지만, 퍼지 최대-최소 신경망을 이용해 특정 집합의 값들을 패턴 공간내의 초상자의 집합으로 변환하고 변환된 초상자들끼리의 인접성을 척도로 단순한 연산을 통한 빠른 특정 집합을 선택하게 된다. 마지막으로 본 논문의 특정 집합 선택 방법을 하지 근전도 신호를 이용한 보행 패턴 분류에 적용해 보고, 그 결과를 기존 여러 특정 집합 선태 방법들과 비교해 봄으로써 제안한 방법의 타당성 및 적용 가능성을 알아본다.

• Korean Journal of Radiology
• /
• v.24 no.1
• /
• pp.51-61
• /
• 2023

Objective: To develop and test a machine learning model for classifying human papillomavirus (HPV) status of patients with oropharyngeal squamous cell carcinoma (OPSCC) using ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG) PET-derived parameters in derived parameters and an appropriate combination of machine learning methods in patients with OPSCC. Materials and Methods: This retrospective study enrolled 126 patients (118 male; mean age, 60 years) with newly diagnosed, pathologically confirmed OPSCC, that underwent ¹⁸F-FDG PET-computed tomography (CT) between January 2012 and February 2020. Patients were randomly assigned to training and internal validation sets in a 7:3 ratio. An external test set of 19 patients (16 male; mean age, 65.3 years) was recruited sequentially from two other tertiary hospitals. Model 1 used only PET parameters, Model 2 used only clinical features, and Model 3 used both PET and clinical parameters. Multiple feature transforms, feature selection, oversampling, and training models are all investigated. The external test set was used to test the three models that performed best in the internal validation set. The values for area under the receiver operating characteristic curve (AUC) were compared between models. Results: In the external test set, ExtraTrees-based Model 3, which uses two PET-derived parameters and three clinical features, with a combination of MinMaxScaler, mutual information selection, and adaptive synthetic sampling approach, showed the best performance (AUC = 0.78; 95% confidence interval, 0.46-1). Model 3 outperformed Model 1 using PET parameters alone (AUC = 0.48, p = 0.047) and Model 2 using clinical parameters alone (AUC = 0.52, p = 0.142) in predicting HPV status. Conclusion: Using oversampling and mutual information selection, an ExtraTree-based HPV status classifier was developed by combining metabolic parameters derived from ¹⁸F-FDG PET/CT and clinical parameters in OPSCC, which exhibited higher performance than the models using either PET or clinical parameters alone.

• Phonetics and Speech Sciences
• /
• v.13 no.2
• /
• pp.57-66
• /
• 2021

This study focuses on the issue of automatic severity classification of dysarthric speakers based on speech intelligibility. Speech intelligibility is a complex measure that is affected by the features of multiple speech dimensions. However, most previous studies are restricted to using features from a single speech dimension. To effectively capture the characteristics of the speech disorder, we extracted features of multiple speech dimensions: voice quality, prosody, and pronunciation. Voice quality consists of jitter, shimmer, Harmonic to Noise Ratio (HNR), number of voice breaks, and degree of voice breaks. Prosody includes speech rate (total duration, speech duration, speaking rate, articulation rate), pitch (F0 mean/std/min/max/med/25quartile/75 quartile), and rhythm (%V, deltas, Varcos, rPVIs, nPVIs). Pronunciation contains Percentage of Correct Phonemes (Percentage of Correct Consonants/Vowels/Total phonemes) and degree of vowel distortion (Vowel Space Area, Formant Centralized Ratio, Vowel Articulatory Index, F2-Ratio). Experiments were conducted using various feature combinations. The experimental results indicate that using features from all three speech dimensions gives the best result, with a 80.15 F1-score, compared to using features from just one or two speech dimensions. The result implies voice quality, prosody, and pronunciation features should all be considered in automatic severity classification of dysarthria.