• Industry Promotion Research
• /
• v.7 no.1
• /
• pp.1-8
• /
• 2022

We investigated various types of machine learning methods that can be applied to censored data. Exploratory data analysis reveals the distribution of each feature, relationships among features. Next, classification problem has been set up where the dependent variable is death_event while the rest of the features are independent variables. After applying various machine learning methods to the data, it has been found that just like many other reports from the artificial intelligence arena random forest performs better than logistic regression. But recently well performed artificial neural network and gradient boost do not perform as expected due to the lack of data. Finally Kaplan-Meier and Cox proportional hazard model have been employed to explore the relationship of the dependent variable (t_i, δ_i) with the independent variables. Also random forest which is used in machine learning has been applied to the survival analysis with censored data.

• Radiation Oncology Journal
• /
• v.10 no.1
• /
• pp.59-67
• /
• 1992

From March 1979 through December 1986, 124 patients with early stage carcinoma of the uterine cervix received curative radiation therapy. According to FIGO classification, 35 patients were stage IB and 89 were stge II A. In stage IB, five year locoregional control, five year disease free survival, and five year overall survival was $79.0\%$, $76.4\%$ and $81.8\%$, respectively. In stage II A, five year locoregional control, five year disease free survival, and five year overall survival were $78.0\%$, $66.8\%$, and $72.1\%$, respectively. To identify prognostic factors, pretreatment parameters including age, ECOG performance status, number of pregnancies, history of diabetes mellitus and hypertension, histology, size and shape of primary tumor, CT findings and blood parameters were retrospectively analyzed in terms of locoregional control, disease free survival and overall survival using univariate analysis and multivariate analysis. In univariate analysis, tumor size on physicai examination and rectal invasion on CT significantly affected locoregional control, disease free survival and overall survival. Parametrial involvement on CT was a significant prognostic factor on locoregional control and disease free survival. Hemoglobin level affected disease free survival and overall survival. Histology and age were significant prognostic factors on locoregional control. In multivariate analysis excluding CT finding, tumor size on physical examination was a significant factor in terms of locoregioal control and overall survival. Hemoglobin level was significant in terms of disease free survival. In multivariate analysis including CT, histology was a prognostic factor on locoregional control and disease free survival. Hemoglobin level and rectal invasion on CT were significant factors on locoregional control.

• Journal of Chest Surgery
• /
• v.34 no.8
• /
• pp.615-620
• /
• 2001

The new stage of metastatic lung cancer based upon resectability, disease-free interval, and the number of pulmonary metastases was proposed in 1998 by Ginsberg, et al. We evaluated the validity of the new staging proposal for pulmonary metastases through the analysis of experiences at Severance Hospital. Material and Method: The cases of 111 patients who underwent resection of metastatic lung cancer during the eleven-year period (1990-2000) were reviewed. Of these patients, 103(92.8%) underwent compete surgical resection. The primary tumor was carcinoma in 60 cases, sarcoma in 46, and others in 5. The disease-free interval(DFI) was 0 to 35 months in 79 cases adn more than 36 months in 32 cases. Single metastasis accounted for 53 cases and multiple lesions for 58 cases. Mean follow-up was 49 months. Result: The actuarial survival after complete metastasectomy was 48.2% at 3 years and 32.6% at 5 years; the corresponding values for incomplete resection were 21.9% at 3 years. The 3-year survival rate(3-YSR) for complete resction was 40.5% and 5-year survival rate(5-YSR) was 30.4% for patients with a DFI less than 36 months, the 3-YSR, 75.8% and 5-YSR, 39.0% for those with a DFI equal or more than 36 moths; 45.8% and 30.5% for single lesions, 50.0% and 34.4% for multiple lesions. The 3-YSR and 5-YSR were 58.5% and 43.8% for stage I patients, 54.0% and 37.4% for stage II, 38.2% and 27.9% for stage III and 21.9% for stage IV. Conclusion: The result of the analysis of new stage of pulmonary metastases showed that the survival rate was different according to stage and there was no statistical significance. We need more experiences and long-term follow up to determine the prognostic factor of metastatic lung cancer surgery.

• Proceedings of KOSOMES biannual meeting
• /
• 2017.04a
• /
• pp.170-170
• /
• 2017

• Journal of Hospice and Palliative Care
• /
• v.15 no.2
• /
• pp.99-107
• /
• 2012

Purpose: Most terminally ill cancer patients die from cancer anorexia-cachexia syndrome. This study evaluated a prognostic role of plasma leptin levels in terminally ill cancer patients. Methods: This study enrolled 69 terminally ill cancer patients who were aged above 20 years old from July 2009 to July 2010. For univariate analysis, an association between leptin levels and patient's characteristics or other variables was examined using Spearman's correlation analysis, Wilcoxon's rank-sum test or Kruskal-Wallis test, as appropriately. For multivariable analysis, Cox's proportional hazard regression model was used to evaluate a clinical significance of plasma leptin levels as a prognostic factor and to determine factors which affect the risk of death in terminally ill cancer patients. Results: A statistically significant positive correlation between plasma leptin levels and survival time was found. Univariate Cox's proportional hazard regression analyses also showed a moderately significant association between plasma leptin levels and survival time. However, after adjusting variables for sex, white blood cell counts, total bilirubin, AST, ALT, albumin and CRP levels, plasma leptin levels were not significantly associated with survival time. Conclusion: No significant association was found between plasma leptin levels and survival time in terminally ill cancer patients. However, this study suggested a prognostic value of plasma leptin levels in gastrointestinal cancer patients.

• Asia-Pacific Journal of Business Venturing and Entrepreneurship
• /
• v.18 no.1
• /
• pp.225-236
• /
• 2023

Policy implications were derived by comparing/analyzing innovative SMEs and general SMEs that obtained innovation certification from 2015 to 2021 in terms of survival period. Work experience, scale (employment, capital and debt size, sales and operating profit) Korean standard industry classification (2 digit) was used to select general SMEs similar to innovative SMEs. Survival period was calculated by defining suspension, closure and overdue equivalent to default as events. As a result of the survival analysis, innovative SMEs showed a 9.8% reduction in the risk of delinquency compared to general SMEs, indicating that the survival period of innovative SMEs was significantly longer. In addition, it was found that the work experience and size (employment, capital) of SMEs had a positive effect on the survival period, but debt had a negative effect on the survival period. This means that the innovation certification system centered on innovation capabilities and future growth potential is a significant indicator in terms of survival period. As a result, it was concluded that the benefits and support policies provided by the innovation certification system need to be more systematic and sophisticated by reflecting the work experience and industry for the actual growth and survival of SMEs.

• Journal of the Korea Society for Simulation
• /
• v.27 no.2
• /
• pp.115-123
• /
• 2018

we analyzed simulation of the effectiveness for one or two DIRCM on a helicopter. The survival rate of helicopter followed increase of the deception rate of DIRCM. When the deception rate was over 70% at 100% detection rate, the survival rate was 10~30% when one DIRCM was installed and the survival rate was 70~80% when two DIRCMs were installed. When the detection rate was over 70% at 100% deception rate the survival rate was 10~30% case of one DIRCM was installed. survival rate was 20~30% when two were installed. Survival rate of 70~90% was observed with one DIRCM when the deception rate and detection rate were 100%, and 100% with two DIRCMs.

• Journal of the military operations research society of Korea
• /
• v.34 no.3
• /
• pp.79-92
• /
• 2008

The purpose of this paper is to design an algorithm for data fusion of sensors data in the helicopter ASE system, using Bayesian Network, which was selected among several knowledge base data fusion methods after consideration and applied to this study. The result of the algorithm analysis shows that Bayesian Network is effective method for solving this problem.

• Radiation Oncology Journal
• /
• v.24 no.2
• /
• pp.110-115
• /
• 2006

Purpose: This study was performed to determine the prognostic factors influencing relapse pattern, overall and disease-free survival in patients treated with postoperative radiotherapy for endometrial carcinoma. Materials and Methods: The records of 54 patients with endometrial adenocarcinoma treated postoperative radiotherapy at Pusan National University Hospital between April 1992 and May 2003 were reviewed retrospectively. Median age of the patients was 55 (range $35{\sim}76$). The distribution by surgical FIGO stages were 63.0% for 0Stage I, 14.8% for Stage II, 22.2% for Stage III. All patients received postoperative external radiotherapy up to $41.4{\sim}54Gy$ (median: 50.4 Gy). Additional Intravaginal brachytherapy was app led to 20 patients (37.0% of all). Median follow-up time was 35 months ($5{\sim}115$ months). Significant factors of this study: histologic grade, Iymphovascular space invasion and myometrial invasion depth were scored (GLM score) and analyzed. Survival analysis was peformed using Kaplan-Meier method. The log-rank test was used for univariate analysis and the Cox regression model for multivariate analysis. Results: 5-year overall and disease-free survival rates were 87.7% and 871%, respectively. Prognostic factors related with overall and disease-free survival were histologic grade, Iymphovascular space invasion and myometrial invasion according to the univariate analysis. According to the multivariate analysis, Iymphovascular space invasion was associated with decreased disease-free survival. GLM score was a meaningful factor affecting overall and disease-free survival (p=0.0090, p=0.0073, respectively) and distant recurrence (p=0.0132), which was the sum of points of histologic grade, Iymphovascular space Invasion and myometrial invasion. Total failure rate was 11% with 6 patients. Relapse sites were 2 para-aortic Iymph nodes, 2 lungs, a supraclavicular Iymph node and a vagina. Conclusion: The prognosos in patients with endometrial carcinoma treated by postoperative radiotherapy was closely related with surgical histopathology. If further explorations confirm the system of prognostic factors in endometrial carcinoma, it will help us to predict the progression pattern and to manage.

• The KIPS Transactions:PartC
• /
• v.12C no.5 s.101
• /
• pp.633-642
• /
• 2005

Ubiquitous computing system is about networked processors, which is constructed with one or more computers interconnected by the networks. However, traditional security solution lacks a Proactive maintenance technique because of its focusing on developing the qualitative detection and countermeasure after attack. Thus, in this paper, we propose a quantitative assessment modeling technique, by which the general infrastructure can be improved and the attacks on a specific infrastructure be detected and protected. First of all, we develop the definition of survivality and modeling technique for quantitative assessment modeling with the static information on the system random information, and attack-type modeling. in addition, the survivality analysis on TCP-SYN attack and code-Red worm attack is performed for validating the proposed technique.