• 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.

• Korean Journal of Remote Sensing
• /
• v.37 no.4
• /
• pp.719-731
• /
• 2021

This study proposes a two-stage hybrid classification model for crop classification using multi-temporal remote sensing images; the model combines feature embedding by using an autoencoder (AE) with a convolutional neural network (CNN) classifier to fully utilize features including informative temporal and spatial signatures. Long short-term memory (LSTM)-based AE (LAE) is fine-tuned using class label information to extract latent features that contain less noise and useful temporal signatures. The CNN classifier is then applied to effectively account for the spatial characteristics of the extracted latent features. A crop classification experiment with multi-temporal unmanned aerial vehicle images is conducted to illustrate the potential application of the proposed hybrid model. The classification performance of the proposed model is compared with various combinations of conventional deep learning models (CNN, LSTM, and convolutional LSTM) and different inputs (original multi-temporal images and features from stacked AE). From the crop classification experiment, the best classification accuracy was achieved by the proposed model that utilized the latent features by fine-tuned LAE as input for the CNN classifier. The latent features that contain useful temporal signatures and are less noisy could increase the class separability between crops with similar spectral signatures, thereby leading to superior classification accuracy. The experimental results demonstrate the importance of effective feature extraction and the potential of the proposed classification model for crop classification using multi-temporal remote sensing images.

• Journal of Korean Institute of Industrial Engineers
• /
• v.40 no.3
• /
• pp.275-282
• /
• 2014

One of the most widely used methods for dimensionality reduction is principal component analysis (PCA). However, the reduced dimensions from PCA do not provide a clear interpretation with respect to the original features because they are linear combinations of a large number of original features. This interpretation problem can be overcome by feature selection approaches that identifying the best subset of given features. In this study, we propose an unsupervised feature selection method based on the geometrical information of PCA loading vectors. Experimental results from a simulation study demonstrated the efficiency and usefulness of the proposed method.

• Journal of Information Processing Systems
• /
• v.12 no.4
• /
• pp.644-660
• /
• 2016

The real-time detection of malware remains an open issue, since most of the existing approaches for malware categorization focus on improving the accuracy rather than the detection time. Therefore, finding a proper balance between these two characteristics is very important, especially for such sensitive systems. In this paper, we present a fast portable executable (PE) malware detection system, which is based on the analysis of the set of Application Programming Interfaces (APIs) called by a program and some technical PE features (TPFs). We used an efficient feature selection method, which first selects the most relevant APIs and TPFs using the chi-square ($KHI^2$) measure, and then the Phi (${\varphi}$) coefficient was used to classify the features in different subsets, based on their relevance. We evaluated our method using different classifiers trained on different combinations of feature subsets. We obtained very satisfying results with more than 98% accuracy. Our system is adequate for real-time detection since it is able to categorize a file (Malware or Benign) in 0.09 seconds.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.8 no.11
• /
• pp.4137-4152
• /
• 2014

This paper investigates the description ability of chaotic feature vector to dynamic textures. First a chaotic feature and other features are calculated from each pixel intensity series. Then these features are combined to a chaotic feature vector. Therefore a video is modeled as a feature vector matrix. Next by the aid of bag of words framework, we explore the representation ability of the proposed chaotic feature vector. Finally we investigate recognition rate between different combinations of chaotic features. Experimental results show the merit of chaotic feature vector for pixel intensity series representation.

• The Korean Journal of Applied Statistics
• /
• v.34 no.2
• /
• pp.153-166
• /
• 2021

A main goal of pharmacogenomics studies is to predict individual's drug responsiveness based on high dimensional genetic variables. Due to a large number of variables, feature selection is required in order to reduce the number of variables. The selected features are used to construct a predictive model using machine learning algorithms. In the present study, we applied several hybrid feature selection methods such as combinations of logistic regression, ReliefF, TurF, random forest, and LASSO to a next generation sequencing data set of 400 epilepsy patients. We then applied the selected features to machine learning methods including random forest, gradient boosting, and support vector machine as well as a stacking ensemble method. Our results showed that the stacking model with a hybrid feature selection of random forest and ReliefF performs better than with other combinations of approaches. Based on a 5-fold cross validation partition, the mean test accuracy value of the best model was 0.727 and the mean test AUC value of the best model was 0.761. It also appeared that the stacking models outperform than single machine learning predictive models when using the same selected features.

• Sleep Medicine and Psychophysiology
• /
• v.18 no.1
• /
• pp.5-9
• /
• 2011

Sleep-related eating disorder (SRED) is a newly recognized parasomnia that describes a clinical condition of compulsive eating under an altered level of consciousness during sleep. Recently, it is increasingly recognized in clinical practice. The exact etiology of SRED is unclear, but it is assumed that SRED might share features of both sleepwalking and eating disorder. There have been also accumulating reports of SRED related to the administration of various psychotropic drugs, such as zolpidem, triazolam, olanzapine, and combinations of psychotropics. Especially, zolpidem in patients with underlying sleep disorders that cause frequent arousals, may cause or augment sleep related eating behavior. A thorough sleep history is essential to recognition and diagnosis of SRED. The timing, frequency, and description of food ingested during eating episodes should be elicited, and a history of concurrent psychiatric, medical, sleep disorders must also be sought and evaluated. Interestingly, dopaminergic agents as monotherapy were effective in some trials. Success with combinations of dopaminergic and opioid drugs, with the addition of sedatives, has also been reported in some case reports.

• Asia pacific journal of information systems
• /
• v.4 no.1
• /
• pp.32-46
• /
• 1994

The present study addresses the research frameworks and directions in the area of accounting information systems (AIS) by conducting a comprehensive survey of 76 A/S doctoral dissertations published between 1980 and 1990. The central research task is to identify the distinctive features of A/S that distinguish it from other disciplines. The procedure of the survey analysis is as follows. First, the A/S dissertations are categorized according to information systems (IS) and accounting. Then, some representative A/S research areas are documented, based on the IS/accounting domain combinations. The paper also attempts to discover the topical trends, supporting bases, and qualitative aspects of the A/S research. The survey reveals a : 1) the existence of economics as a supporting discipline, 2) the topical trend consistent with the evolution of IS, but restricted by the specific combinations of its parent domains, and 3) the need for intellectual rigor in the A/S research.

• Clinical Psychopharmacology and Neuroscience
• /
• v.16 no.4
• /
• pp.376-382
• /
• 2018

We reviewed clinical studies investigating the pharmacological treatment of major depressive episodes (MDEs) with mixed features diagnosed according to the dimensional criteria (more than two or three [hypo]manic symptoms+principle depressive symptoms). We systematically reviewed published randomized controlled trials on the pharmacological treatment of MDEs with mixed features associated with mood disorders, including major depressive disorder (MDD) and bipolar disorder (BD). We searched the PubMed, Cochrane Library, and ClinicalTrials.gov databases through December 2017 with the following key word combinations linked with the word OR: (a) mixed or mixed state, mixed features, DMX, mixed depression; (b) depressive, major depressive, MDE, MDD, bipolar, bipolar depression; and (c) antidepressant, antipsychotic, mood stabilizer, anticonvulsant, treatment, medication, algorithm, guideline, pharmacological. We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. We found few randomized trials on pharmacological treatments for MDEs with mixed features. Of the 36 articles assessed for eligibility, 11 investigated MDEs with mixed features in mood disorders: six assessed the efficacy of antipsychotic drugs (lurasidone and ziprasidone) in the acute phase of MDD with mixed features, although four of these were post hoc analyses based on large randomized controlled trials. Four studies compared antipsychotic drugs (olanzapine, lurasidone, and ziprasidone) with placebo, and one study assessed the efficacy of combination therapy (olanzapine+fluoxetine) in the acute phase of BD with mixed features. Pharmacological treatments for MDEs with mixed features have focused on antipsychotics, although evidence of their efficacy is lacking. Additional well-designed clinical trials are needed.

• Proceedings of the KSME Conference
• /
• 2008.11a
• /
• pp.1833-1836
• /
• 2008

Recently as the micro surface features become higher and diverse in their shapes, the releasing of the molded features becomes more crucial for manufacturing of the micro patterned products. The higher aspect ratio of the features or more complex shape of the features results in larger releasing force, elongation or cohesive failure of the features during the releasing. Another issue would be the uniformity of the released surface features after molding, especially for applications with large area surface. The micro patterned optical film, one of typical applications for micro surface features, consists of two layers, the thermoplastic base film and the micro formed UV resin layer. Therefore two interfaces are typically involved during the forming of this micro featured film; one is between the base film and the UV resin and another is between the resin and the pattern master. To improve the releasing of the molded surface features, the adhesive characteristic was investigated at these two interfaces. A PET film was used as a base film and two UV curable resins with different surface energy were prepared for different adhesiveness. Also the two different pattern masters were employed; one is made from brass-copper alloy and fabricated with PMMA. The adhesiveness at each interface was measured for some combinations of these base film, UV resins and the masters and the effect of this adhesiveness on the releasing was investigated.