Jea-Bok Ryu

A Short Consideration of Binomial Confidence Interval

The interval estimation for binomial proportion has been treated practically as well as theoretically for a long time. In this paper we compared the properties of major confidence intervals and summarized current issues for coverage probability and interval length which are the criteria of evaluation for confidence interval. Additionally, we examined the three topics which were considered in using the binomial confidence interval in the field. And finally we discussed the future studies for a low binomial proportion.

The Influence of Extreme Value in Binomial Confidence Interval

이항비율에 대한 구간추정에 다양한 신뢰구간들이 사용된다. 그러나 대부분의 신뢰구간들은 모비율 p가 0이나 1에 근사할 때 포함확률이 신뢰수준(또는 명목수준, 1-alpha)을 크게 벗어난다. 이는 극단적인 관찰값의 영향 때문이다. Vollset (1993), Agresti와 Coull (1998), Newcombe (1998), Brown 등 (2001) 등은 극단값의 조정을 통해서 이러한 문제를 해결하는 방법들을 제시하였다. 본 연구에서는 극단값들이 이항비율에 대한 신뢰구간에 어느 정도 영향을 미치는 지를 6개의 신뢰구간들에 대해서 수치적으로 비교해 보았다.

Penalized regression models for patent keyword analysis

Technology analysis is important work in management of technology. Most companies make plans for research and development (R&D) policy, new product development, or technological innovation using the results of technology analysis. In this paper, we propose a methodology of technology analysis using penalized regression models. We analyze the patent keywords extracted from the patent documents using ridge regression, least absolute shrinkage and selection operator, elastic net, and random forest. In addition, to show how our research could be applied to real problem efficiently, we carry out a case study of Apple technology. Our study contributes to perform R&D planning in technology management.

Confidence Interval for Sensitive Binomial Attribute : Direct Question Method and Indirect Question Method

We discuss confidence intervals for sensitive binomial attributes obtained by a direct question method and indirect question method. The Randomized Response Technique(RRT) by Warner (1965) is an indirect question method that uses a randomization device to reduce the response burden of respondents. We used the mean coverage probability (MCP), root mean squared error (RMSE), and mean expected width (MEW) to compare the confidence intervals by the two methods. The numerical comparisons indicated found that the MEW of RRT is too large and the RRT is so conservative that the MCP exceeds a nominal level( ); therefore, it is necessary to complement these problem in order to increase the utility of the indirect question method.

Imputation Methods for the Population and Housing Census 2000 in Korea

We proposed imputation strategies for the Population and Housing Census 2000 in Korea. The total area of floor space and marital status which have relatively high non-response rates in the Census are considered to develope the effective missing value imputation procedures. The Classification and Regression Tree(CART) is employed to construct the imputation cells for hot-deck imputation, as well as to predict missing value by model-based approach. We compare three imputation methods which include CART model-based imputation, hot-deck imputation based on CART and logical hot-deck imputation proposed by The Korea National Statistical Office. The results suggest that the proposed hot-deck imputation based on CART is very efficient and strongly recommendable.

A Sampling Design for Health Index Survey

We propose a new sampling design for the 2001 Health Index Survey at Seoul. In this stratified two-stage sampling design, the ED(enumeration district) of 2000 Population and Housing Census is used as primary sampling unit and the Gu is used as stratification variable in order to obtain the sub-domain estimate for 25 Gus as well as population estimate for Seoul. The sample EDs are systematically selected after the Eds are ordered by location and property to obtain a representative sample. And also, the imputation methods for item nonresponses are suggested.