• The Korean Journal of Applied Statistics
• /
• v.29 no.1
• /
• pp.231-242
• /
• 2016

X-13-ARIMA (a popular time series analysis software) provides $3{\times}3$, $3{\times}5$, $3{\times}9$, $3{\times}15$ moving average filters for seasonal adjustment. However, there has been questions on their performance and the need for new filters is a constant topic due to Korean economic time series often containing higher irregularity and more various seasonality than other countries. In this study, two newly developed seasonal moving average filters, $3{\times}7$ and $3{\times}11$, are introduced. New filters were implemented in X-13-ARIMA and applied to 15 economic time series to demonstrate their suitability and reliability. The result shows that some series are more stable when using new seasonal moving average filters. More accurate time series analyses would be possible if newly proposed filters are used together with existing filters.

• The Korean Journal of Financial Management
• /
• v.16 no.1
• /
• pp.171-191
• /
• 1999

시간의 흐름에 걸친 주가시계열의 행동양식에 대한 연구에서는 선형성, 비선형성, 장기기억, 항상성분 등에 대한 명확한 결론을 내리고 있지 못한 실정이다. 주가 시계열과정을 설명하고 예측하기 위한 여러 모형들에 대한 실증연구에는 설명력과 예측력을 완벽하게 갖추고 있지 못하고 있다는 증거들이 제시되고 있다. 계절적 변동을 주가시계열에 적용하지 않는 관계로 이와 같은 결과가 발생할 가능성이 존재한다. 분기별 종합주가지수의 수익률에 계절적 단위근이 존재하고 있음이 실증분석을 통하여 밝혀졌다. 이 시계열에서는 계절적 단위근을 제거하기 위하여서는 제4계 시차 작용소가 적절한 필터임이 인정되었다. 월별 종합주가지수의 수익률에서도 계절적 단위근이 존재하고 있다. 따라서 제12계 시차 작용소를 사용하여 계절적 단위근을 제거하여야 할 것이다. 분기별 수익률에는 제4차 시차 작용소를, 월별수익률에서는 제12차 시차 작용소를 필터로 사용하여 이 시계열들을 차분화하고 이 차분화를 통하여 계절적 단위근을 제거한 후에 이 시계열들의 시계열적 성질과 특성을 탐구해야 할 것이다. 이 과정을 통할 때 시계열 과정에 대한 계량경제학적 모형에 대한 정확한 추론이 가능하게 된다.

• The Korean Journal of Applied Statistics
• /
• v.27 no.1
• /
• pp.133-146
• /
• 2014

The United States Census Bureau released a new version of X-13ARIMA-SEATS that integrates X-12-ARIMA with TRAMO-SEATS. This paper compares a seasonal adjusted series from X-13ARIMA-SEATS and those from X-12-ARIMA. An X11 filter and SEATS filter were used for the X-13ARIMA-SEATS. The result of the comparison suggests that seasonal adjusted series using X-13ARIMA-SEATS with the X11 filter are similar to those of X-12-ARIMA.

• The Korean Journal of Applied Statistics
• /
• v.23 no.6
• /
• pp.997-1021
• /
• 2010

This paper compares the two most widely used seasonal adjustment methods: the X-12-ARIMA and TRAMO-SEATS procedures. The basic features of these methods are discussed and compared in both their theoretical and empirical aspects. In doing so, the X-13A-S program is used to reevaluate their applicability to Korean macroeconomic data by considering possible structural breaks in the series. The finding is that both methods provide very reliable and stable estimates of seasonal factors and seasonally adjusted data. As for the empirical comparisons, TRAMO-SEATS appears to outperform X-12-ARIMA, although the results are somewhat mixed depending on the comparison criteria used and on the series under analysis. In particular, the performance of TRAMO-SEATS turns out to compare more favorably when seasonal adjustment is carried out to each sub-samples (by taking possible structural breaks into account) than when the whole sample period is used. The result suggests that as the model-based TRAMO-SEATS has a considerable theoretical appeal, some features of TRAMO-SEATS should further be incorporated into X-12-ARIMA until a standard and integrated procedure is reached by combining the theoretical coherence of TRAMO-SEATS and the empirical usefulness of X-12-ARIMA.

• Proceedings of the Korea Air Pollution Research Association Conference
• /
• 1988.12a
• /
• pp.33-34
• /
• 1988

• Journal of the Korean Geographical Society
• /
• v.43 no.1
• /
• pp.36-51
• /
• 2008

As the global warming has influenced on various sectors including agriculture, forestry, fisheries and health, it is essential to project more accurate future climate for an assessment of climate change impact and adaptation strategy. This study examines spatial distribution of onset dates and durations of season decomposed by applying a lowpass filtering using observed 30-year (1971-2000) data and projected 2090s data based on the IPCC SRES A1B emission scenario in South Korea. In general, the distributions of spring and winter onset date are affected by latitudes, topography and proximity to oceans. However, onset dates of summer and autumn are a little affected by proximity to oceans and topography than by latitudes. In the 2090s (2091-2100), the onset dates of spring begin about 40 days earlier and the onset dates of summer begin 25-30 days earlier as compare with present time. On the other hand, the onset dates of winter begin about 50 days later in the southern and eastern coastal area and in the southern inland. The onset dates of autumn begin about 20 days later. In the 2090s, summer duration is longer and winter duration is shorter as compare with present time at southern and eastern coastal area.

• Journal of Korean Society of Forest Science
• /
• v.13 no.1
• /
• pp.1-24
• /
• 1971

In order to elucidate the feasibility to identify the plant species through the photographic tone this study was made. In this study the author made photographs of Juniperus chinensis L., and Populus glandulosa Uyeki, with the panchromatic film using either yellow filter or red filter in different seasons respectively. The author analyzed the value of tone variation at the level of stereoscopic view of the same photographs by using tone scale and Automatic Micro-Photo-Densitometer. The results obtained are summarized as follows: A. Tone scale reading 1. The tone of Populus glandulosa Uyeki was darker than that of Juniperus chinensis L. in a photograph. 2. Regardless of the tree species, tone of photographs obtained with yellow filter was darker than that with red filter. 3. Along the progresses of seasons, the photographic tone was changed. That is, from the spring to the summer it showed darker and than, from the summer to winter it changed lighter. 4. During winter and spring the discrimination between the both species of trees can be easily made by stereoscopic view whether there are leaves on the tree or not rather than by tone observation. 5. Regardless of tree species, variation of tone due to the age was noticed. The older trees have darker tone than the younger one. 6. It is recognized that the yearly difference depends on insolation quantities. 7. The highest reflex light-waves were in between $600m{\mu}$ and $660m{\mu}$ for both species of trees. B. Density reading 1. For the density reading, there was the same tendency as in tone scale reading. 2. The changes of peaks of the scanning curves with Populus glandulosa Uyeki is smoother and takes place in lower position than with Juniperus chinensis L. 3. The scanning curves on the 20th May was smoothest, and change of peaks increased gradually according to the season progresses. 4. In relation to the types of filter the photographs with yellow filter showed less changes of peaks then that with red filter.

• Journal of the Korean Geographical Society
• /
• v.42 no.6
• /
• pp.835-850
• /
• 2007

This study examines long-term trends of onset dates and durations of season decomposed by applying a lowpass filtering using observed 80-year(1921-2000) data and projected 2040s and 2090s data based on the IPCC SRES A1B emission scenario in South Korea. As recent climate changes on seasonal cycles in South Korea, the onset dates of spring and summer began earlier after 1970s. But onset dates of autumn and winter were delayed. Winter duration was more shortened during the post-1990 period. Summer duration is longer and winter duration is shorter at southern and eastern coastal area. The different of summer duration and winter duration in coastal area was longer than over the inland. The reduction in winter duration in South Korea agrees with results in overall global warming trends as a climate change signal. Future model output data predict that winter will disappear in Gangneung, Busan and Mokpo in the 2090s

• Proceedings of the KIEE Conference
• /
• 2000.07b
• /
• pp.1445-1447
• /
• 2000

본 논문에서는 철도 수요예측 문제의 유형을 목적에 따라 3가지로 분류하였고, 최근 철도자원을 재고관리 차원에서 접근하고자 하는 시각에 따라 분배모형으로써 적응필터를 사용하는 방법의 타당성에 대해 설명하였다. 또 철도 승객수요의 주요 특징을 분석하였으며, 철도 승객수요 예측의 요구사항 및 방법론을 대규모 재고관리 시스템의 일반적 요구사항에 따라 정리하였다. 영향인자에 대한 분석으로 요일별 계절변동 지수를 정량적으로 산정하였다. 적응필터를 이용한 철도 승객수요 예측의 예제를 제시하였으며, 예측에의 정확성에 대한 비교를 제시하였다.

• The Bulletin of The Korean Astronomical Society
• /
• v.36 no.1
• /
• pp.66.1-66.1
• /
• 2011

천체망원경을 이용한 측광관측은 주로 CCD를 이용하여 이루어지고 있으며, CCD를 이용한 측광관측은 CCD를 구성하는 픽셀마다 양자효율이 일정하지 않기 때문에 Flat 관측을 이용하여 보정을 하고 있다. 한편, Flat 관측은 주로 빈 하늘을 찍는 Sky Flat 방법을 이용하고 있으나, Sky Flat을 관측하는 방법은 계절에 따라 시작 시간이 다르고, 제한적인 관측 시간 내에 관측자의 판단으로 이루어지고 있는 실정이다. 특히 제한적인 관측시간은 관측자의 상당한 신중함을 요구하고 있어 노출시간 조절 및 시작 시간을 지나게 되면 Sky Flat 영상을 얻지 못하는 경우도 있다. 이러한 제약 조건에서 효율적인 Flat관측을 하기위해 AutoFlat 프로그램을 개발하였다. AutoFlat 프로그램은 5초 노출로 빈 하늘을 찍으며 관측된 이미지에서 정해진 하늘의 밝기가 되면 자동으로 노출 시간을 계산하고, 각각의 필터별로 관측을 하는 동시에 적절한 이름을 주어 자동으로 저장하게 되어 있다. 이번 연구에서 개발한 AutoFlat 프로그램은 관측자가 Flat 관측을 진행하지 않아도 안정적이면서 효율적인 Flat 관측을 수행하여 앞으로 CCD를 이용한 측광 관측에서 유용하게 사용될 것으로 기대된다.