• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.23 no.6
• /
• pp.496-501
• /
• 2011

The missing data patterns of tidal elevation data in Korean coasts are analysed and provided. The missing interval of the data is displayed for all stations using the missing data indicator matrix in order to identify the overall missing pattern. The spatial and temporal missing rates are also estimated. The total missing rate of tidal elevation data is low. However, most of the missing is mainly derived from just 1 or 2 specific stations. The autocorrelation function of the consecutive missing interval data also shows that the missing interval occurs randomly.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2004.04a
• /
• pp.499-501
• /
• 2004

웹 로그, 바이오정보학 둥 여러 분야에서 다양한 형태의 결측치가 발생하여 학습 데이터를 희소하게 만든다. 결측치는 주로 전처리 과정에서 조건부 평균이나 나무 모형과 같은 기본적인 Imputation 방법을 이용하여 추정된 값에 의해 대체되기도 하고 일부는 제거되기도 한다. 특히, 결측치 비율이 매우 크게 되면 기존의 결측치 대체 방법의 정확도는 떨어진다. 또한 데이터의 결측치 비율이 증가할수록 사용 가능한 Imputation 방법들의 수는 극히 제한된다. 이러한 문제점을 해결하기 위하여 본 논문에서는 Vapnik의 Support Vector Regression을 데이터 전처리 과정에 알맞게 변형한 Support Vector Regression을 제안하여 이러한 문제점들을 해결하였다. 제안 방법을 통하여 결측치의 비율이 상당히 큰 희소 데이터의 전처리도 가능하게 되었다. UCI machine learning repository로부터 얻어진 데이터를 이용하여 제안 방법의 성능을 확인하였다.

• The Korean Journal of Applied Statistics
• /
• v.15 no.1
• /
• pp.33-43
• /
• 2002

Various kinds of estimation methods have been developed for imputation of categorical missing data. They include category method, logistic regression, and association rule. In this study, we propose two fusions algorithms based on both neural network and voting scheme that combine the results of individual imputation methods. A Mont-Carlo simulation is used to compare the performance of these methods. Five factors used to simulate the missing data pattern are (1) input-output function, (2) data size, (3) noise of input-output function (4) proportion of missing data, and (5) pattern of missing data. Experimental study results indicate the following: when the data size is small and missing data proportion is large, modal category method, association rule, and neural network based fusion have better performances than the other methods. However, when the data size is small and correlation between input and missing output is strong, logistic regression and neural network barred fusion algorithm appear better than the others. When data size is large with low missing data proportion, a large noise, and strong correlation between input and missing output, neural networks based fusion algorithm turns out to be the best choice.

• The Korean Journal of Applied Statistics
• /
• v.29 no.3
• /
• pp.549-557
• /
• 2016

Consecutive missing values are likely to occur in long climate data due to system error or defective equipment. Furthermore, it is difficult to impute missing values. However, these complicated problems can be overcame by imputing missing values with reference time series. Reference time series must be composed of similar time series to time series that include missing values. We performed a simulation to compare three missing imputation methods (the adjusted normal ratio method, the regression method and the IDW method) to complete the missing values of time series. A comparison of the three missing imputation methods for the daily mean temperatures at 14 climatological stations indicated that the IDW method was better thanx others at south seaside stations. We also found the regression method was better than others at most stations (except south seaside stations).

• Proceedings of the Korea Water Resources Association Conference
• /
• 2006.05a
• /
• pp.1546-1550
• /
• 2006

한강 하구부의 3개 지점에서 수중 계류방식으로 약 5개월에 걸쳐 탁도를 관측하였다. 이 과정에서 관측기기의 한계로 인해 탁도 자료의 결측치가 발생하였고, 이를 효율적으로 보완하기 위해 새로운 결측치 보완기법을 개발하였다. 개발된 기법, 일명 면적비법은 시계열 자료가 단일주기와 상이한 진폭을 갖는다는 가정하에, 각 사이클의 면적비율을 통해 결측치를 보완하는 방법이다. 면적비법과 기존의 최소제곱법을 검증하기 위해 결측치가 없는 정상적인 자료에 적용해 보면, 두 방법 모두 첨두치를 약간 과소 산정하는 경향이 있었다. 하지만 면적비법의 경우, 원자료의 총 면적과 보완자료의 총 면적간의 차이가 거의 없었다. 이 방법들을 한강 하구부에서 관측된 탁도자료에 적용해 본 결과, 면적비법은 합리적으로 결측치를 보완하는 반면, 최소제곱법은 보완자료의 총면적이 원자료에 비해 작아지는 오류가 발생하였다. 따라서 최소제곱법에 비해 면적비법이 결측치 보완에 더 우수한 결과를 제공함을 알 수 있었다. 본 연구에서 개발한 면적비법은 주기성이 뚜렷한 시계열자료의 결측치 보완에 유용하게 쓰일 수 있으리라 기대된다.

• The Korean Journal of Applied Statistics
• /
• v.13 no.1
• /
• pp.197-206
• /
• 2000

For estimating missing cells in contingency table, we suggest an iterative method which extends IPF (Iterative Proportional Fitting) method. The suggested m~thod is not restricted by the number and the location of missing cells, and does not distort the given quasi-independency.

• Journal of Korean Society of Transportation
• /
• v.24 no.7 s.93
• /
• pp.27-40
• /
• 2006

The traffic information is provided, which based on the volume of traffic, speed, occupancy collected through the currently operating Vehicle Detector System(VDS). In addition to the trend in utilization fold of traffic information is increasing gradually with the applied various fields and users. Missing data in Vehicle detector data means series of data transmitted to controller without specific property. The missing data does not have a data property, so excluded at the whole data Process Hence, increasing ratio of missing data in VDS data inflicts unreliable representation of actual traffic situation. This study presented the imputation process due out which applied the methodologies that utilized adjacent stations reference and historical data utilize about missing data. Applied imputation process methodologies to VDS data or SeoHaeAn/Kyongbu Expressway, currently operation VDS, after processes at missing data ratio of an option. Imputation process held presented to per lane-30seconds-period, and morning/afternoon/daily time scope ranges classified, and analyzed an error of imputed data preparing for actual data. The analysis results, an low error occurred relatively in the results of the imputation process way that utilized a historical data compare with adjacent stations reference methods.

• Journal of Korea Water Resources Association
• /
• v.39 no.4 s.165
• /
• pp.289-298
• /
• 2006

Turbidity had been measured during five months In Han River estuary at three sites. In this process, missing data occurred due to gauge imitation of the turbidity sensor. A filling method for the missing turbidity data was newly developed in this study. Under the assumption of the time series data with unique period and different amplitudes, the new method can fill the missing data based on the area ratio of each cycle. And the new method was verified through the data set having no missing data. There were little differences between gross area of the original data and that of the revised data by the new method though values of peak were underestimated. As a result, missing turbidity data observed at Han River estuary could be appropriately filled using the new filling method.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.14 no.6
• /
• pp.789-792
• /
• 2004

In various fields as web mining, bioinformatics, statistical data analysis, and so forth, very diversely missing values are found. These values make training data to be sparse. Largely, the missing values are replaced by predicted values using mean and mode. We can used the advanced missing value imputation methods as conditional mean, tree method, and Markov Chain Monte Carlo algorithm. But general imputation models have the property that their predictive accuracy is decreased according to increase the ratio of missing in training data. Moreover the number of available imputations is limited by increasing missing ratio. To settle this problem, we proposed statistical learning theory to preprocess for missing values. Our statistical learning theory is the support vector regression by Vapnik. The proposed method can be applied to sparsely training data. We verified the performance of our model using the data sets from UCI machine learning repository.

• Journal of Korean Society of Transportation
• /
• v.34 no.5
• /
• pp.437-448
• /
• 2016

Some previous researches suggested a transit trip destination inference method by constructing trip chains with incomplete(missing destination) smart card dataset obtained on the entry fare control systems. To explore the feasibility of the transit trip destination inference method, the transit trip chains are constructed from the pre-paid smart card tagging data collected in Busan on October 2014 weekdays by tracing the card IDs, tagging times(boarding, alighting, transfer), and the trip linking distances between two consecutive transit trips in a daily sequences. Assuming that most trips in the transit trip chains are linked successively, the individual transit trip destination zones are inferred as the consecutive linking trip's origin zones. Applying the model to the complete trips with observed OD reveals that about 82% of the inferred trip destinations are the same as those of the observed trip destinations and the inference error defined as the difference in distance between the inferred and observed alighting stops is minimized when the trip linking distance is less than or equal to 0.5km. When applying the model to the incomplete trips with missing destinations, the overall destination missing rate decreases from 71.40% to 21.74% and approximately 77% of the destination missing trips are the single transit trips for which the destinations can not be inferable. In addition, the model remarkably reduces the destination missing rate of the multiple incomplete transit trips from 69.56% to 6.27%. Spearman's rank correlation and Chi-squared goodness-of-fit tests showed that the ranks for transit trips of each zone are not significantly affected by the inferred trips, but the transit trip distributions only using small complete trips are significantly different from those using complete and inferred trips. Therefore, it is concluded that the model should be applicable to derive a realistic transit trip patterns in cities with the incomplete smart card data.