• Journal of IKEEE
• /
• v.24 no.3
• /
• pp.766-776
• /
• 2020

Recently, as interest in machine learning (ML) has increased and research using ML has become active, it is becoming more important to find an optimal hyperparameter combination for various ML models. In this paper, among various hyperparameters, we focused on ML optimizers, and measured and compared the performance of major optimizers using various datasets. In particular, we compared the performance of nine optimizers ranging from SGD, which is the most basic, to Momentum, NAG, AdaGrad, RMSProp, AdaDelta, Adam, AdaMax, and Nadam, using the MNIST, CIFAR-10, IRIS, TITANIC, and Boston Housing Price datasets. Experimental results showed that when Adam or Nadam was used, the loss of various ML models decreased most rapidly and their F1 score was also increased. Meanwhile, AdaMax showed a lot of instability during training and AdaDelta showed slower convergence speed and lower performance than other optimizers.

• Survey Research
• /
• v.6 no.1
• /
• pp.83-98
• /
• 2005

Large scale survey database may contain some erroneous data or missing data. Incomplete or erroneous data may be produced in the process of data collection or data capture. Since erroneous data can cause some bias and inconsistency, data editing, which is the procedure for detecting and adjusting individual errors in data records, is a very important work in statistical survey. In this paper, we introduce an editing process for the housing price survey to enhance discussions on that topic. We explain how to decide some appropriate edit rules and show some related data. Furthermore, we describe input editing procedures which is appropriate for on-line survey and how to find and eliminate erroneous data through output editing.

• Proceedings of the Korean Institute Of Construction Engineering and Management
• /
• 2007.11a
• /
• pp.704-709
• /
• 2007

The purpose of this study is to present both problems and improvement plans on the bidding system of the Korea National Housing Corporation, which orders apartment houses professionally. For data gathering, this study used data from electronic procurement system from 2001 to 2005. The research method of this study is to analyze the data categorized into four types: changes of ordering and contract price ration, the lowest-pricing bidding system, and permanent projects by grades. As a result of the analysis, the amount of ordering is biased in favor of particular level. Finally, this study present following improvement plans: 1) screening system and the lowest-pricing bidding system need to be supplemented, 2) the contract price ratio and the amount of ordering should be controlled, and 3) the improvement of PQ and the introduction of the list of Qualified companies by types need to be presented.

• Korea Real Estate Review
• /
• v.28 no.3
• /
• pp.65-77
• /
• 2018

The purpose of this study is to analyze the effect of the educational environment on the prices of studio apartments, known as officetels, in Korea. Since the revision of relevant laws in 2010, they have served as substitutes for residential purposes in areas suffering from housing shortage, especially where the educational environment is a significant factor. To assess the relation between the educational environment and rental rates, the hedonic price model and artificial neural networks were implemented. The national assessment of the academic performance of middle and high schools that were closest to each officetel, and the ratio of students going to special-purpose schools and private high schools were considered as independent variables. Research findings indicated that the positive effect of the dependent variable increased as the value of educational environment-related variables increased. This result could be utilized as a functional index for housing providers after considering educational environments.

• Land and Housing Review
• /
• v.14 no.4
• /
• pp.39-46
• /
• 2023

This study explores the interrelationship between 'KRX Construction' and 'Business Survey Index'. KRX Construction is a leading economic indicator of construction industry, implying the potential interdependence with BSI Construction. Previous papers have investigated the relationship between the released stock price index and BSI. Using Granger causality tests, this study investigates how the BSI Construction is associated with the trend and noise-trading components of KRX Construction, respectively. The decomposition of KRX Construction of trend and noise-trading is based on the state-space model. The results document unilateral Granger causalities from released KRX Construction, trend component, noise-trading component to BSI Construction. In sum, this study demonstrates that construction company CEOs view stock price index as a leading economic indicator.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.9 no.3
• /
• pp.377-386
• /
• 2014

This study analyzed how alternative investment goods would affect a market in a neighboring shopping area in order to provide parties involved in the investment market of this neighboring shopping area with standards which would help them when they try to make a reasonable determination. The study estimated forms and explanation power of the effects of a bid rate of a neighboring shopping area, and came up with those results as follows. Increases in the representative macro economic indicators, the composite stock price index and the fluctuation rate of land price, including the real estate business would have a positive influence on the market of the neighboring shopping area as playing a circumstantial evidence of market recovery and yet, the increase in interest rate, the alternative investment goods, would reduce the relative price-earnings ratio which would, eventually, negatively affect the charm of the investment in the market of the neighboring shopping area. The study, now, understands that housing with a feature of consumers' goods and neighboring shopping area with a feature of investment goods would not have great concern with each other as they are observed to be two different markets from an aspect of interactionism.

• Journal of the Korean Home Economics Association
• /
• v.18 no.4
• /
• pp.65-73
• /
• 1980

The purpose of this study is ti find out a desirable way to stability and improvement of household economy by studying the changes of consumption level and consumption pattern of urban salary and wage earners' households during the years from 1970 to 1978. For this study, "Annual Report on the Family Income and Expenditure survey" (Published by the Bureau of Statistics, Economics Planning Board) has been used as basic material, and the methods of analysis used here are the time series analysis. We have gained the results as follows: 1) From 1970 to 198, the total income level increased at the rate of 416.2% in nominal price, but only 74.4% in reql price, while the total expenditure level showed 338.5% increase in nominal price, but its real increased proved only 418.2% in consideration of inflation. APC decreased from 95.1%(in 1970) to 80.7%(in 1978). 2) As for the expenditure pattern for the above mentioned nine years, the rate of food expenditure increased until 1975 under the price influence, but it trended to decease there after on . The rate of housing expenditure showed a gradual increase while that of fuel and light expenditure was on the decrease. The rate of clothing expenditure had been on the decease until 1974 but it began to increase gradually thereafter on. The trend of miscellaneous expenditures was irregularly up and down, educational expences being the first rank among them, Non-living expenditure had been constant until 1974 but it decreased a little after that. From the results it was found that the consumption level of the salary and wage earners' household in all cities from 1970 to 1978 was not practically improved because of rise in prices, nor was the Engel's coefficient and the rate of miscellaneous expenditure changed distinctively. However, as the successive decrease of APC suggests the possibility of economic development, we must try to put stress on economy in consumption and on encouraging. This will help run our household economy in safety and stability.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.15 no.3
• /
• pp.1318-1326
• /
• 2014

The change of concept about real estate marketing and the solution about unsold houses are getting important nowadays. This paper is analyzing the solution of unsold houses as location environment, residential environment, macro environment, and marketing, and research question is which elements are important to real users. In addition, this paper is analyzing on impact relationship in terms of real users' buying intention about unsold houses and satisfaction of price of selling a house (price of selling a house, rise in housing value) by PLS structural equation. To sum up, first, macro environment, residential environment, location environment is statistically valued on the satisfaction of price of selling a house. However, marketing strategic is not impacted on the satisfaction of price of selling. Second, macro environment, marketing strategic, satisfaction of selling a house is statistically valued on buying intention.

• Journal of the Korean Data and Information Science Society
• /
• v.28 no.2
• /
• pp.327-337
• /
• 2017

In this study we compared three models for small area estimation, Fay-Herriot, Hierarchical Bayses model and spatio-temporal model about charter, monthly rent price index. Charter, monthly rent price of Korea are important issue in these days. Because housing type rapidly changes from self to charter and monthly rent. The accuracy of the estimation was checked on four scales, that is ARB, ASRB, AAB, ASD. In this result, the spatio-temporal model among applied models has most optimal scales about small area estimation of charter and monthly rent index.

• Korean Journal of Construction Engineering and Management
• /
• v.7 no.5
• /
• pp.94-104
• /
• 2006

The changed Korean government law associated with the public apartment housing supply, so called the $^{\circ}{\AE}$Apartment Sales Price $Cap^{\circ}{\phi}$, requires new system for estimating construction cost in order to set appropriate price. A model apartment project was carefully designed and its construction cost were analyzed in many different ways. Based on the analyses outcomes, 1,028,000 Won/m2 (excluding cost for underground parking lot) is the most appropriate Price Cap for a smaller than $85{\beta}{\geq}$ apartment unit. Further, it was revealed that the price have to be adjusted reflecting such factors as underground size; structural system; external complex quality; and consumer preferences. Findings from this study will enable the Korean government to realize faster and better application of the related laws. The methodology for obtaining appropriate apartment construction cost will also benefit for future researchers.