• Korean Journal of Remote Sensing
• /
• v.36 no.6_2
• /
• pp.1591-1604
• /
• 2020

This study is a study on classifying land cover by applying high-resolution satellite images to deep learning algorithms and verifying the performance of algorithms for each spatial object. For this, the Fully Convolutional Network-based algorithm was selected, and a dataset was constructed using Kompasat-3 satellite images, land cover maps, and forest maps. By applying the constructed data set to the algorithm, each optimal hyperparameter was calculated. Final classification was performed after hyperparameter optimization, and the overall accuracy of DeeplabV3+ was calculated the highest at 81.7%. However, when looking at the accuracy of each category, SegNet showed the best performance in roads and buildings, and U-Net showed the highest accuracy in hardwood trees and discussion items. In the case of Deeplab V3+, it performed better than the other two models in fields, facility cultivation, and grassland. Through the results, the limitations of applying one algorithm for land cover classification were confirmed, and if an appropriate algorithm for each spatial object is applied in the future, it is expected that high quality land cover classification results can be produced.

• Proceedings of the Korea Technology Innovation Society Conference
• /
• 1999.11c
• /
• pp.137-141
• /
• 1999

• Journal of the Korea Convergence Society
• /
• v.11 no.10
• /
• pp.45-51
• /
• 2020

Recently, many games provide data related to the users' game play, and there have been a few studies that predict opponent move by combining machine learning methods. This study predicts opponent move using match data of a real-time strategy game named ClashRoyale and a multi-label classification based on machine learning. In the initial experiment, binary card properties, binary card coordinates, and normalized time information are input, and card type and card coordinates are predicted using random forest and multi-layer perceptron. Subsequently, experiments were conducted sequentially using the next three data preprocessing methods. First, some property information of the input data were transformed. Next, input data were converted to nested form considering the consecutive card input system. Finally, input data were predicted by dividing into the early and the latter according to the normalized time information. As a result, the best preprocessing step was shown about 2.6% improvement in card type and about 1.8% improvement in card coordinates when nested data divided into the early.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.26 no.3
• /
• pp.190-195
• /
• 2016

A technology is continuously developed in a rapidly changing global market. A company requires an appropriate R&D strategy for adapting to this environment. That is, the technologies owned by the company needs to be thoroughly analyzed to improve its competitiveness. Alternatively, technology classification using IPC codes is carried out recently in an objective and quantitative way. International Patent Classification, IPC is an internationally specified classification system, so it is helpful to conduct an objective and quantitative patent analysis of technology. In this study, all of the patents owned by company C are investigated and a matrix representing IPC codes of each patent is created. Then, a structured association map of the patents is made through association rules mining based on Confidence. The association map can be used to inspect the current situation of a company about patents. It also allows highly associated technologies to be clustered. Using the association map, this study analyzes the technologies of company C and how it changes with time. The strategy for future technologies is established based on the result.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1996.10a
• /
• pp.177-181
• /
• 1996

Systematic classification system for standardization in telecommunication is essential to the standardization R&D strategy. This paper suggests a new reference model for development of work area and classification scheme related to the telecommunications standardization : Cubic and matrix approach. Standardization Work Areas(SWAs) that are upper level of the reference model are classified by its main role and function reflecting the market trends and user needs. Standardization expertise is lower level scheme, which can be regarded as the different possible layers of standardization to be applied to each one of the SWAs grouped under upper level scheme. A new reference model consists of two planes that are SWAs plane and Standardization layer plane. Finally the reference model for classification of SWAs in telecommunication mapping onto matrix table that row and column are defined by SWAs and standardization layer respectively.

• Journal of Information Technology Applications and Management
• /
• v.19 no.1
• /
• pp.25-36
• /
• 2012

Social Commerce recently attracted the attention of academic and industry researchers. Social Commerce aims to make a transactional environment which is beneficial to three parties-social commerce service provider, buyer and seller by way of using the platform of SNS. As Social Commerce is a new technology issue, there is no existing conceptual framework, e.g. appropriate classification the business types, that help to understand the nature of Social Commerce. This study suggests one classification framework and tries to verify whether it works.

• Journal of The Korean Astronomical Society
• /
• v.29 no.spc1
• /
• pp.125-126
• /
• 1996

Beijing-Arizona-Taipei-Connecticut (BATC) survey is a long term project to map the spectral energy distribution of various objects using 15 intermediate band filters and aims to cover about 450 sq degrees of northern sky. The SED information, combined with image structure information, is used to classify objects into several stellar and galaxy categories as well as QSO candidates. In this paper, we present a preliminary setup of robust data reduction procedure recently developed at NCU and also briefly discuss general classification scheme: redshift estimate, and automatic detection of variable objects.

• Proceedings of the Korean Statistical Society Conference
• /
• 2005.11a
• /
• pp.141-146
• /
• 2005

Ensemble method has been known as one of the most powerful classification tools that can improve prediction accuracy. Ensemble method also has been understood as ‘perturb and combine’ strategy. Many studies have tried to develop ensemble methods by improving perturbation. In this paper, we propose two new ensemble methods that improve combining, based on the idea of pattern matching. In the experiment with simulation data and with real dataset, the proposed ensemble methods peformed better than bagging. The proposed ensemble methods give the most accurate prediction when the pruned tree was used as the base learner.

• Journal of Information Technology Applications and Management
• /
• v.10 no.4
• /
• pp.51-63
• /
• 2003

This study designs and implements a new approach to the classification of e-mail requests from customer based on machine learning techniques. The work on building an electronic mall classifier can be cast into the framework of text classification, since an e-mail is a viewed as a document, and judgement of interest is viewed as a class level given to the e-mail document. It is also implemented an e-mall based automated response system that integrate with Call Center in a practical use.

• Proceedings of the Korea Database Society Conference
• /
• 1997.10a
• /
• pp.215-269
• /
• 1997

Outline $\textbullet$ Introduction $\textbullet$ Multimedia - Types of Data - Motivation - Key issue - Hardware Products - Application Areas $\textbullet$ Agents - Rationale for Agents - Sedentary vs. Mobile - Functional Categories - Application Areas $\textbullet$ Data Mining - 2-D Framework for Data Mining Tools - Classification of Tool - Application Areas - Learning Methodologies ＊ Case Based Reasoning ＊ Neural Networks ＊ Statistical Learning： Orthogonal Arrays ＊ Multi-strategy Learning $\textbullet$ Case Study - Finbot $\textbullet$ Conclusion