• Journal of the Korea Society of Computer and Information
• /
• v.28 no.7
• /
• pp.19-27
• /
• 2023

In the process of collecting social reviews, a number of noise reviews irrelevant to a given search keyword can be included in the search results. To filter out such reviews, machine learning can be used. However, if the number of reviews is insufficient for a target place to be analyzed, filtering accuracy can be degraded due to the lack of training data. To resolve this issue, we propose a supervised learning method to improve accuracy of the noise review filtering for the places with insufficient reviews. In the proposed method, training is not performed by an individual place, but by a group including several places with similar characteristics. The classifier obtained through the training can be used for the noise review filtering of an arbitrary place belonging to the group, so the problem of insufficient training data can be resolved. To verify the proposed method, a noise review filtering model was implemented using LSTM and BERT, and filtering accuracy was checked through experiments using real data collected online. The experimental results show that the accuracy of the proposed method was 92.4% on the average, and it provided 87.5% accuracy when targeting places with less than 100 reviews.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• v.18 no.2
• /
• pp.143-156
• /
• 2014

This paper aims to give a quick view on denoising without comprehensive details. Denoising can be understood as removing unwanted parts in signals and images. Noise incorporates intrinsic random fluctuations in the data. Since noise is ubiquitous, denoising methods and models are diverse. Starting from what noise means, we briefly discuss a denoising model as maximum a posteriori estimation and relate it with a variational form or energy model. After that we present a few major branches in image and signal processing; filtering, shrinkage or thresholding, regularization and data adapted methods, although it may not be a general way of classifying denoising methods.

• Journal of Korean Society for Quality Management
• /
• v.46 no.3
• /
• pp.379-396
• /
• 2018

Purpose: The purpose of this study was to propose a standard form of noise filter drawing, which is widely used in the field of weapon systems, to define the essential specification which is not standardized at the existing drawings and to employ the reference for manufacturing and qualification test. Methods: To do that, it identifies the improvements to analyze 19 noise filter drawings, studies the principle of noise filtering to find the essential part of the spec. and reviews the MIL-STD-220C and KS IEC 60939, which is the widely used as a formal spec. pertaining to the noise filter. Throughout the study, a standard form is developed to standardize the attenuation characteristics and find eight crucial metrics to specify the noise filter. Results: The results of this study are as follows; The standardized drawings can be used as the qualification test of the item and the manufacturing and engineering documents to make the noise filter, which are also applicable to the reference document to review KDS(Korea Defense Standard) drawings.

• Korean Business Review
• /
• v.19 no.1
• /
• pp.81-92
• /
• 2006

This study investigates the relation between institutional investor's net purchase and the volatility of KOSPI. Some portion of volatility in stock prices comes from noise trading of irrational traders. Observed volatility may be defined as the sum of the portion caused by information arrival, fundamental volatility, and the portion caused by noise trading, transitory volatility. This study decomposes the observed volatility into fundamental volatility and transitory volatility using Kalman filtering method. Most studies investigates the effect on the observed volatility. In contrast to other studies, this study investigates the effect on the fundamental volatility and transitory volatility individually. Estimation results show that institutional investor's net purchase was not significantly related to all kinds of volatility(observed volatility, fundamental volatility and transitory volatility). This means that institutional investor's net purchase did not increase noise trading.

• Korean Business Review
• /
• v.20 no.2
• /
• pp.15-27
• /
• 2007

This paper analyzed the relationship between Housing Purchase Price volatility of Seoul and Housing Purchase Price volatility of local large city. Other studies investigates the effect on the observed volatility Observed volatility consists of fundamental volatility and transitory volatility. Fundamental volatility is caused by information arrival and transitory volatility is caused by noise trading. Fundamental volatility is trend component and is modelled as a random walk with drift. Transitory volatility is cyclical component and is modelled as a stationary process. In contrast to other studies, this study investigates the effect on the fundamental volatility and transitory volatility individually. Observed volatility is estimated by GJR GARCH(1,1) model. We find that GJH GARCH model is superior to GARCH model and good news is more remarkable effect on volatility than bad news. This study decomposes the observed volatility into fundamental volatility and transitory volatility using Kalman filtering method. The findings in this paper is as follows. The correlation between Seoul housing price volatility and Busan housing price volatility is high. But, the correlation between Seoul and Daejeon is low. And the correlation between Daejeon and Busan is low. As a distinguishing feature, the correlation between fundamental volatilities is high in the case of all pairs. But, the correlation between transitory volatilities turns out low. The reason is as follows. When economic information arrives, Seoul, Daejeon, and Busan housing markets, all together, are affected by this information.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.8
• /
• pp.65-72
• /
• 2020

Big data analysis is widely used in applications such as finance and communication, whose market size is growing rapidly every year. Nevertheless, it is rarely used by SMEs (small and medium-sized enterprises) since the existing services are not fully customized for them while being offered at high price. To resolve this, we develop and propose a new platform to provide big data analysis services specialized for SMEs in this paper. First, we compare existing work discussing social big data analysis, and extract service features necessary to help their marketing effectively. Then, we present a prototype system implementing the extracted features, and discuss technical issues needed to develop a complete system which are obtained from the prototype implementation.

• Geophysics and Geophysical Exploration
• /
• v.23 no.3
• /
• pp.117-130
• /
• 2020

Noises can distort acquired geophysical data, leading to their misinterpretation. Potential noises sources include anthropogenic activity, natural phenomena, and instrument noises. Conventional denoising methods such as wavelet transform and filtering techniques, are based on subjective human investigation, which is computationally inefficient and time-consuming. Recently, many researchers attempted to implement neural networks to efficiently remove noise from geophysical data. This study aims to review and analyze different types of neural networks, such as artificial neural networks, convolutional neural networks, autoencoders, residual networks, and wavelet neural networks, which are implemented to remove different types of noises including seismic, transient electromagnetic, ground-penetrating radar, and magnetotelluric surveys. The review analyzes and summarizes the key challenges in the removal of noise from geophysical data using neural network, while proposes and explains solutions to the challenges. The analysis support that the advancement in neural networks can be powerful denoising tools for geophysical data.