• Asian Pacific Journal of Cancer Prevention
• /
• 제14권1호
• /
• pp.159-163
• /
• 2013

Background: This study used Surveillance, Epidemiology and End Results (SEER) pancreatic cancer data to identify predictive models and potential socio-economic disparities in pancreatic cancer outcome. Materials and Methods: For risk modeling, Kaplan Meier method was used for cause specific survival analysis. The Kolmogorov-Smirnov's test was used to compare survival curves. The Cox proportional hazard method was applied for multivariate analysis. The area under the ROC curve was computed for predictors of absolute risk of death, optimized to improve efficiency. Results: This study included 58,747 patients. The mean follow up time (S.D.) was 7.6 (10.6) months. SEER stage and grade were strongly predictive univariates. Sex, race, and three socio-economic factors (county level family income, rural-urban residence status, and county level education attainment) were independent multivariate predictors. Racial and socio-economic factors were associated with about 2% difference in absolute cause specific survival. Conclusions: This study s found significant effects of socio-economic factors on pancreas cancer outcome. These data may generate hypotheses for trials to eliminate these outcome disparities.

• Journal of the Korean Data and Information Science Society
• /
• 제26권3호
• /
• pp.749-754
• /
• 2015

This article presents Bayesian approach to regression splines with knots on a grid of equally spaced sample quantiles of the independent variables under functional measurement error model.We consider small area model by using penalized splines of non-linear pattern. Specifically, in a basis functions of the regression spline, we use radial basis functions. To fit the model and estimate parameters we suggest a hierarchical Bayesian framework using Markov Chain Monte Carlo methodology. Furthermore, we illustrate the method in an application data. We check the convergence by a potential scale reduction factor and we use the posterior predictive p-value and the mean logarithmic conditional predictive ordinate to compar models.

• 통합자연과학논문집
• /
• 제4권4호
• /
• pp.266-272
• /
• 2011

Focal adhesion kinase (FAK) is a potential target for the treatment of primary cancers as well as prevention of tumor metastasis. To understand the structural and chemical features of FAK inhibitors, we report comparative molecular field analysis (CoMFA) for the series of 7H-pyrrolo(2,3-d)pyrimidines. The CoMFA models showed good correlation between the actual and predicted values for training set molecules. Our results indicated the ligand-based alignment has produced better statistical results for CoMFA ($q^2$ = 0.505, $r^2$ = 0.950). Both models were validated using test set compounds, and gave good predictive values of 0.537. The statistical parameters from the generated 3D-QSAR models were indicated that the data are well fitted and have high predictive ability. The contour map from 3D-QSAR models explains nicely the structure-activity relationships of FAK inhibitors and our results would give proper guidelines to further enhance the activity of novel inhibitors.

• Journal of information and communication convergence engineering
• /
• 제21권1호
• /
• pp.82-89
• /
• 2023

Various machine-learning models may yield high predictive power for massive time series for time series prediction. However, these models are prone to instability in terms of computational cost because of the high dimensionality of the feature space and nonoptimized hyperparameter settings. Considering the potential risk that model training with a high-dimensional feature set can be time-consuming, we evaluate a feature-importance-based feature selection method to derive a tradeoff between predictive power and computational cost for time series prediction. We used two machine learning techniques for performance evaluation to generate prediction models from a retail sales dataset. First, we ranked the features using impurity- and Local Interpretable Model-agnostic Explanations (LIME) -based feature importance measures in the prediction models. Then, the recursive feature elimination method was applied to eliminate unimportant features sequentially. Consequently, we obtained a subset of features that could lead to reduced model training time while preserving acceptable model performance.

• Geomechanics and Engineering
• /
• 제10권3호
• /
• pp.269-284
• /
• 2016

Simplified techniques based on in situ testing methods are commonly used to assess seismic liquefaction potential. Many of these simplified methods were developed by analyzing liquefaction case histories from which the liquefaction boundary (limit state) separating two categories (the occurrence or non-occurrence of liquefaction) is determined. As the liquefaction classification problem is highly nonlinear in nature, it is difficult to develop a comprehensive model using conventional modeling techniques that take into consideration all the independent variables, such as the seismic and soil properties. In this study, a modification of the Multivariate Adaptive Regression Splines (MARS) approach based on Logistic Regression (LR) LR_MARS is used to evaluate seismic liquefaction potential based on actual field records. Three different LR_MARS models were used to analyze three different field liquefaction databases and the results are compared with the neural network approaches. The developed spline functions and the limit state functions obtained reveal that the LR_MARS models can capture and describe the intrinsic, complex relationship between seismic parameters, soil parameters, and the liquefaction potential without having to make any assumptions about the underlying relationship between the various variables. Considering its computational efficiency, simplicity of interpretation, predictive accuracy, its data-driven and adaptive nature and its ability to map the interaction between variables, the use of LR_MARS model in assessing seismic liquefaction potential is promising.

• Journal of Computing Science and Engineering
• /
• 제5권3호
• /
• pp.183-196
• /
• 2011

There has recently been a surge of interest in relational database mining that aims to discover useful patterns across multiple interlinked database relations. It is crucial for a learning algorithm to explore the multiple inter-connected relations so that important attributes are not excluded when mining such relational repositories. However, from a data privacy perspective, it becomes difficult to identify all possible relationships between attributes from the different relations, considering a complex database schema. That is, seemingly harmless attributes may be linked to confidential information, leading to data leaks when building a model. Thus, we are at risk of disclosing unwanted knowledge when publishing the results of a data mining exercise. For instance, consider a financial database classification task to determine whether a loan is considered high risk. Suppose that we are aware that the database contains another confidential attribute, such as income level, that should not be divulged. One may thus choose to eliminate, or distort, the income level from the database to prevent potential privacy leakage. However, even after distortion, a learning model against the modified database may accurately determine the income level values. It follows that the database is still unsafe and may be compromised. This paper demonstrates this potential for privacy leakage in multi-relational classification and illustrates how such potential leaks may be detected. We propose a method to generate a ranked list of subschemas that maintains the predictive performance on the class attribute, while limiting the disclosure risk, and predictive accuracy, of confidential attributes. We illustrate and demonstrate the effectiveness of our method against a financial database and an insurance database.

• 대한산업공학회지
• /
• 제40권2호
• /
• pp.240-249
• /
• 2014

This study aims to identify an assessment system based on multiple patent indices that can predict the likelihood of success in the commercialization of a patented technology in advance. In addition, we examine the effectiveness of our predictive model in identifying valuable technologies early on. We analyzed 3,063 secondary battery technologies patented in the US over the past 10 years. Our analysis identified 22 of the 25 most promising patented technologies, corresponding with the top 50% of industry-patented technologies that directly and indirectly succeeded in commercialization. These results support our claim that it is possible to identify attributes for the assessment of patent commercial potential to a significant degree. Our system presents a useful assessment index in the forecasting and determination of potential commercial success of patented technologies.

• Journal of Korean Neurosurgical Society
• /
• 제42권2호
• /
• pp.92-96
• /
• 2007

Objective : The authors have speculated that metastatic brain lesions from renal cell carcinoma (RCC) show diverse radiological patterns and tumor responses after Gamma knife surgery (GKS), and have hypothesized that these can be predicted from tumor radiological characteristics. The goal of the current study was to identify the radiological characteristics of RCC brain metastases and the predictors of initial radiosurgical response after GKS. Methods : A retrospective analysis was performed on 48 lesions in 18 patients with RCC brain metastasis treated by GKS. The radiological characteristics of these lesions in magnetic resonance images (MRI) were classified into 3 categories according to enhancement patterns in T1-weighted images and signal intensity characteristics in T2-weighted images. Responses to GKS were analyzed according to these categories, and in addition, other potential predictive factors were also evaluated. Results : MRI findings in the three categories were diverse, though numbers of the lesion were comparable. At 2-month MRI follow-ups after GKS, response rate was 54% and the local tumor control rate 83%. T2 signal intensity was found to be the principal predictive factor of response to GKS, namely negative predictive factor. Other variables such as age, sex, tumor volume, dose, duration from initial diagnosis to GKS, and previous systemic therapies failed to show significant relationships with treatment response by multivariate analysis. Conclusion : Careful evaluation of the radiological characteristics of brain metastases from RCC is important prior to GKS because MRI heterogeneity has predictive value in terms of determining initial tumor response.

• Corrosion Science and Technology
• /
• 제19권6호
• /
• pp.302-309
• /
• 2020

This study aimed to predict the pitting corrosion characteristics of AL-6XN super-austenitic steel using multiple linear regression. The variables used in the model are degree of sensitization, temperature, and pH. Experiments were designed and cyclic polarization curve tests were conducted accordingly. The data obtained from the cyclic polarization curve tests were used as training data for the multiple linear regression model. The significance of each factor in the response (critical pitting potential, repassivation potential) was analyzed. The multiple linear regression model was validated using experimental conditions that were not included in the training data. As a result, the degree of sensitization showed a greater effect than the other variables. Multiple linear regression showed poor performance for prediction of repassivation potential. On the other hand, the model showed a considerable degree of predictive performance for critical pitting potential. The coefficient of determination (R2) was 0.7745. The possibility for pitting potential prediction was confirmed using multiple linear regression.

• 경영정보학연구
• /
• 제25권3호
• /
• pp.59-81
• /
• 2023

한국의 노인(65세 이상) 자살 문제는 점차 심각해지고 있는 추세이다. 급격한 인구 고령화 흐름에 따라 이러한 고령층의 자살 추세가 더욱 가속화될 것으로 추정되고 있어, 노인 자살을 예방하고 감소시키는 것이 개인 뿐만 아니라 중요한 사회적 과제로 대두되고 있다. 따라서 본 연구는 한국 노인들을 대상으로 자살 생각의 원인 요인을 파악하고 예측 모델을 개발하는 것을 목적 한다. 본 연구는 한국복지패널조사에서 제공하는 7개년의 패널 데이터를 활용하였으며 자살의 대인 관계 이론(interpersonal theory of suicide)과 사회 해체 이론(social disorganization theory)을 바탕으로 노인 자살의 잠재 원인 요인들을 선정한다. 다음으로 노인의 자살 생각에 대한 원인 요인 파악을 위해 패널 로짓 모형 분석을 진행하고 노인 자살 생각의 예측 모델 개발을 위해 딥 러닝과 머신 러닝 알고리즘을 활용한다. 본 연구는 계량 모형 분석을 통해 검증한 주요 원인 요인들을 활용하여 노인 자살을 예방할 수 있는 구체적인 노인 복지 정책 수립에 기여하고자 한다. 본 연구에서 제시된 예측 모델은 자살 고위험군 노인을 선별하고 관리할 수 있는 방안 마련의 기반을 제공한다. 또한 본 연구는 혼합방법론의 시너지를 보였다는 점에서 학술적 시사점을 가진다.