• Journal of the Korean Data and Information Science Society
• /
• v.20 no.2
• /
• pp.301-309
• /
• 2009

Multidimensional scaling(MDS) is a statistical multivariate analysis technique that is often used in information visualization for exploring similarities or dissimilarities in data. In order to analyse and visualize data, MDS measures the dissimilarities between objects and uses them or their mean if they are repeatedly measured. When there exist outliers or when the variation of data is too large, we can hardly get reliable results on the research using MDS. In this paper, we consider the MDS based on bootstrap method when the variation of data is large. Standardized residual sum of squares is considered as measuring goodness-of-fit of the model. A real data analysis is include to examine our approach.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.150-150
• /
• 2022

This study investigates the roles of decomposition methods on high accuracy in daily rainfall prediction from light gradient boosting machine (LightGBM) algorithm. Here, empirical mode decomposition (EMD) and singular spectrum analysis (SSA) methods were considered to decompose and reconstruct input time series into trend terms, fluctuating terms, and noise components. The decomposed time series from EMD and SSA methods were used as input data for LightGBM algorithm in two hybrid models, including empirical mode-based light gradient boosting machine (EMDGBM) and singular spectrum analysis-based light gradient boosting machine (SSAGBM), respectively. A total of four parameters (i.e., temperature, humidity, wind speed, and rainfall) at a daily scale from 2003 to 2017 is used as input data for daily rainfall prediction. As results from statistical performance indicators, it indicates that the SSAGBM model shows a better performance than the EMDGBM model and the original LightGBM algorithm with no decomposition methods. It represents that the accuracy of LightGBM algorithm in rainfall prediction was improved with the SSA method when using multivariate dataset.

• The Korean Journal of Applied Statistics
• /
• v.12 no.1
• /
• pp.97-107
• /
• 1999

프로파일 분석은 반복측정 자료를 분석하는데 있어서 널리 사용되는 다변량 분석모형이다. 프로파일 분석에서는 처리 그룹간의 비교와 반응 프로파일의 평행성 검정을 위해서 4가지 검정통계량이 널리 사용되고 있다. 이들 검정통계량은 Wilks의 통계량($\Lambda$), Pillai's Trace 통계량(V), Hotelling-Lawley Trace 통계량(U), Roy's Maximum Root 통계량($\Theta$ )이다. 그 동안 이들 통계량들을 비교하기 위한 여러 연구가 있었지만 주로 일반적인 다변량 분산분석 모형에 근거한 비교였다. 본 논문에서는 자료가 반복측정 자료이고 우리의 관심이 프로파일 분석에 있을 때에 이 4가지 통계량의 비교에 초점을 맞추었다.

• The Korean Journal of Applied Statistics
• /
• v.9 no.1
• /
• pp.45-52
• /
• 1996

In this paper, we describe the basic properties of the parallel coordinate plots and propose a dynamic parallel coordinate plot. This dynamic parallel coordinate plot can be used as a dynamic graphics tool for multivariate data analysis.

• Asian Pacific Journal of Cancer Prevention
• /
• v.15 no.21
• /
• pp.9405-9410
• /
• 2014

Purpose: To evaluate the prognostic impact of peritoneal washing cytology in patients with endometrial and ovarian cancers. Materials and Methods: We retrospectively identified 86 individuals with ovarian carcinomas, ovarian borderline tumors and endometrial adenocarcinomas. The patients had been treated at Shahid Sadoughi Hospital and Ramazanzadeh Radiotherapy Center, Yazd, Iran between 2004 and 2012. Survival differences were determined by Kaplan-Meier analysis. Multivariate analysis was performed using the Cox regression method. A p<0.05 value was considered statistically significant. Results: There were 36 patients with ovarian carcinomas, 4 with borderline ovarian tumors and 46 with endometrial carcinomas. The mean age of the patients was $53.8{\pm}15.2years$. In patients with ovarian carcinoma the overall survival in the negative cytology group was better than the patients with positive cytology although this difference failed to reach statistical significance (p=0.30). At 0 to 50 months the overall survival was better in patients with endometrial adenocarcinoma and negative cytology than the patients with positive cytology but then it decreased (p=0.85). At 15 to 60 months patients with FIGO 2009 stage IA-II endometrial andocarcinoma and negative peritoneal cytology had a superior survival rate compared to 1988 IIIA and positive cytology only, although this difference failed to reach statistical significance(p=0.94). Multivariate analysis using Cox proportional hazards model showed that stage and peritoneal cytology were predictors of death. Conclusions: Our results show good correlation of peritoneal cytology with prognosis in patients with ovarian carcinoma. In endometrial carcinoma it had prognostic importance. Additional research is warranted.

• The Korean Journal of Applied Statistics
• /
• v.31 no.6
• /
• pp.721-732
• /
• 2018

Latent class analysis (LCA) is an important tool to explore unseen latent groups in multivariate categorical data. In practice, it is important to select a suitable set of variables because the inclusion of too many variables in the model makes the model complicated and reduces the accuracy of the parameter estimates. Dean and Raftery (Annals of the Institute of Statistical Mathematics, 62, 11-35, 2010) proposed a headlong search algorithm based on Bayesian information criteria values to choose meaningful variables for LCA. In this paper, we propose a new variable selection procedure for LCA by utilizing posterior probabilities obtained from each fitted model. We propose a new statistic to measure the adequacy of LCA and develop a variable selection procedure. The effectiveness of the proposed method is also presented through some numerical studies.

• Journal of Gastric Cancer
• /
• v.5 no.2
• /
• pp.113-119
• /
• 2005

Purpose: Some controversies exist over the prognostic values of lymphatic, venous, and neural invasion in patients with gastric cancer. This study was conducted to confirm the prognostic values of these histopathologic factors in gastric cancer patients who received a gastrectomy. Materials and Methods: Data for clinicopathologic factors and clinical outcomes were collected retrospectively from the medical records of 1,018 gastric cancer patients who received a gastrectomy at Yeungnam University Medical Center between January 1995 and December 1999. A statistical analysis was done using the SPSS program for Windows (Version 10.0, SPSS Inc., USA). The Kaplan-Meier method was used for the survival analysis. Prognostic factors were analyzed by using a multivariate analysis with Cox proportional hazard regression model. Results: Ages ranged from 21 to 79 (median age, 56). A univariate analysis revealed that age, tumor size, location, gross type, depth of invasion, extent of gastrectomy or lymph node dissection, lymph node metastasis, distant metastasis, lymphatic invasion, venous invasion, neural invasion, pathologic stage, histologic type, and curability of surgery had statistical significance. Among these factors, lymph node metastasis, curability of surgery, neural invasion, lymphatic invasion, and depth of invasion were found to be independent prognostic factors by using a multivariate analysis. Venous invasion showed no prognostic value in the multivariate analysis. Conclusion: Neural invasion and lymphatic invasion are useful parameters in determining a prognosis for gastric cancer patients.

• Journal of Nuclear Fuel Cycle and Waste Technology(JNFCWT)
• /
• v.12 no.4
• /
• pp.275-286
• /
• 2014

This study evaluated the evolution stage and origin of chemical components of 12 boreholes at crystalline bedrock using multivariate statistical and groundwater quality analyses. Groundwater types are mostly belonged to Na(Ca)-$HCO_3$ and Ca-$HCO_3$ types, indicating that directly reaction of cation exchange ($Ca^{2+}{\rightarrow}Na^+$) prevailed. The degree of groundwater evolution is included the range from low to intermediate stage based on field and laboratory analytical conditions. As a result of multivariate statistical analysis, a typical indicator of groundwater contamination, $NO_3$-, has the positive correlation with $Na^+$ and $Cl^-$. The origin of sea spary ($Cl^-$) has the positive correlation with $Na^+$, $SO{_4}^{2-}$, $Mg^{2+}$, and $K^+$, while not correlation with $Ca^{2+}$, $Fe^{2+}$, $HCO_3{^-}$, $F^-$, and $SiO_2$. The concentration of $Cl^-$ and $NO_3{^-}$ belongs to general quality of groundwater and not exceeds over the Korean standard for drinking water. And the negative values of saturation index of minerals are calculated with chemical components in groundwater. Therefore, most of chemical components of groundwater in the study area are originated from natural process between rock and groundwater, whereas some of components are derived from sea spary and anthropogenic sources related to agricultural activities.

• Journal of the Korean Society of Safety
• /
• v.39 no.1
• /
• pp.104-113
• /
• 2024

Unlike in general manufacturing process, safety management in laboratory-based research area is complicated because the latter generally involves trying untested methods or handling unusual substances in small amounts. Laboratory accidents in South Korea have recently shown an increasing trend. Unfortunately, statistics on such accidents are not officially published by any domestic public agencies. In this study, multivariate analysis was performed on the relationships between variables to develop effective strategies for preventing laboratory accidents. A Cross-Tabulation Analysis of accident-related factors in 179 accident cases revealed that the laboratory type, accident type, and unsafe-act type are all statistically significant, whereas the unsafe condition and management factors differ with the statistical criteria. Furthermore, the results of a Multiple-Correspondence Analysis showed that accidents can be divided largely into three groups having different accident causes and injury types; this confirms the necessity of different strategies to prevent accidents of each type. The findings also reveal differences between the distribution of accident types mentioned in the accident case collection books and actual reported cases. This suggests that an official statistical system administered by a public institution would be necessary for effective prevention of laboratory accidents.

• Journal of Korean Neurosurgical Society
• /
• v.57 no.4
• /
• pp.266-270
• /
• 2015

Objective : Chronic subdural hematoma (CSDH) is one of the most frequent problems encountered in neurosurgery. Although burr-hole trephination is widely performed to treat CSDH, the incidence rate of recurrent CSDH is still 2-37%. The goal of this study is to determine the risk factors that affect recurrent CSDH. Methods : A total of 182 patients were included in this study who underwent burr-hole trephination. The clinical factors and radiographic features between the recurrence and the no recurrence groups were analyzed to find the parameters related to the postoperative recurrence of CSDH. Results : For the recurrence of CSDH that occurred in 25 patients (13.7%), among various risk factors, pre and postoperative midline displacements, which are more than 10 mm (p=0.000), and preoperative hemiparesis (p=0.026) had contributed to recurrent CSDH with statistical significance by univariate analysis. Unilateral CSDH were more frequently related to recurrent CSDH (16.3%), although it was not a statistical significant result (p=0.052). Furthermore, preoperative midline displacement only had statistical meaning for the recurrence of CSDH by multivariate analysis. Conclusion : This study indicates that the midline displacement on the preoperative computed tomography scan is the only independent predictor for the recurrence of CSDH.