scholarly journals Computing the Hazard Ratios Associated With Explanatory Variables Using Machine Learning Models of Survival Data

2021 ◽  
pp. 364-378
Author(s):  
Sameer Sundrani ◽  
James Lu
Keyword(s):  
Machine Learning ◽  
Survival Data ◽  
Proportional Hazards ◽  
Cox Proportional Hazards ◽  
Training Data ◽  
Data Sets ◽  
Data Set ◽  
Explanatory Variables ◽  
Cancer Data ◽  
Pan Cancer

PURPOSE The application of Cox proportional hazards (CoxPH) models to survival data and the derivation of hazard ratio (HR) are well established. Although nonlinear, tree-based machine learning (ML) models have been developed and applied to the survival analysis, no methodology exists for computing HRs associated with explanatory variables from such models. We describe a novel way to compute HRs from tree-based ML models using the SHapley Additive exPlanation values, which is a locally accurate and consistent methodology to quantify explanatory variables’ contribution to predictions. METHODS We used three sets of publicly available survival data consisting of patients with colon, breast, or pan cancer and compared the performance of CoxPH with the state-of-the-art ML model, XGBoost. To compute the HR for explanatory variables from the XGBoost model, the SHapley Additive exPlanation values were exponentiated and the ratio of the means over the two subgroups was calculated. The CI was computed via bootstrapping the training data and generating the ML model 1,000 times. Across the three data sets, we systematically compared HRs for all explanatory variables. Open-source libraries in Python and R were used in the analyses. RESULTS For the colon and breast cancer data sets, the performance of CoxPH and XGBoost was comparable, and we showed good consistency in the computed HRs. In the pan-cancer data set, we showed agreement in most variables but also an opposite finding in two of the explanatory variables between the CoxPH and XGBoost result. Subsequent Kaplan-Meier plots supported the finding of the XGBoost model. CONCLUSION Enabling the derivation of HR from ML models can help to improve the identification of risk factors from complex survival data sets and to enhance the prediction of clinical trial outcomes.

scholarly journals Prognostic Impact of a Lymphocyte Activation-Associated Gene Signature in GBM based on Transcriptome Analysis

2020 ◽  
Author(s):  
Yujia Lan ◽  
Erjie Zhao ◽  
Xinxin Zhang ◽  
Xiaojing Zhu ◽  
Linyun Wan ◽  
...  
Keyword(s):  
Proportional Hazards ◽  
Lymphocyte Activation ◽  
Risk Groups ◽  
Gene Signature ◽  
Cox Proportional Hazards ◽  
Training Data ◽  
Data Sets ◽  
Data Set ◽  
Cox Proportional Hazards Regression ◽  
Proportional Hazards Regression

Abstract Background Glioblastoma multiforme (GBM) is the most aggressive primary central nervous system malignant tumor that has poor prognosis. Lymphocyte activation played important roles in cancer and therapy. Objective To identify an efficient lymphocyte activation-associated gene signature that could predict the progression and prognosis of GBM. Methods We used univariate Cox proportional hazards regression and stepwise regression algorithm to develop a lymphocyte activation-associated gene signature in the training data set (TCGA, n = 525). Then, the signature was validated in two data sets, including GSE16011 (n = 150) and GSE13041 (n = 191) by the Kaplan Meier method. Univariate and multivariate Cox proportional hazards regression models were used to adjust for clinicopathological factors. Results In the training data set, we identified a lymphocyte activation-associated gene signature (TCF3, IGFBP2, TYRO3 and NOD2), which classified the patients into high-risk and low-risk groups with significant differences in overall survival (median survival 15.33 months vs 12.57 months, HR = 1.55, 95% CI = 1.28-1.87, logrank test P < 0.001). In the other two data sets, this signature showed similar prognostic values. Further, univariate and multivariate Cox proportional hazards regression models analysis indicated that the signature was an independent prognostic factor for GBM patients.Moreover, we found that there were differences in lymphocyte activity between the high- and low-risk groups of GBM patients among all data sets. Furthermore, by longitudinal analysis, the lymphocyte activationassociated gene signature could significantly predict the survival of patients with some features, such as IDH-wildtype patients and patients with radiotherapy. In addition, the signature could improve prognostic power of age. Conclusions In summary, our results suggested that the lymphocyte activation-associated gene signature is a promising factor for the survival of patients, which may be helpful for diagnosis, prognosis, and treatment of GBM patients.

Facial Attractiveness: Beauty and the Machine

2006 ◽  
Vol 18 (1) ◽  
pp. 119-142 ◽  
Author(s):  
Yael Eisenthal ◽  
Gideon Dror ◽  
Eytan Ruppin
Keyword(s):  
Machine Learning ◽  
Facial Attractiveness ◽  
Training Data ◽  
Data Sets ◽  
Learning Context ◽  
Data Set ◽  
Attractiveness Rating ◽  
Facial Beauty ◽  
Facial Images

This work presents a novel study of the notion of facial attractiveness in a machine learning context. To this end, we collected human beauty ratings for data sets of facial images and used various techniques for learning the attractiveness of a face. The trained predictor achieves a significant correlation of 0.65 with the average human ratings. The results clearly show that facial beauty is a universal concept that a machine can learn. Analysis of the accuracy of the beauty prediction machine as a function of the size of the training data indicates that a machine producing human-like attractiveness rating could be obtained given a moderately larger data set.

COMPARISON OF PROPORTIONAL HAZARDS MODELING, ADDITIVE HAZARDS MODELING AND PROPORTIONAL INTENSITY MODELING WHEN APPLIED TO REPAIRABLE SYSTEM RELIABILITY

1995 ◽  
Vol 02 (01) ◽  
pp. 23-34 ◽  
Author(s):  
DAVE WIGHTMAN ◽  
TONY BENDELL
Keyword(s):  
System Reliability ◽  
Proportional Hazards ◽  
Data Sets ◽  
Reliability Data ◽  
Data Set ◽  
Explanatory Variables ◽  
Additive Hazards ◽  
Proportional Intensity ◽  
Modeling Techniques ◽  
Selection Of

In an Industrial Reliability setting a number of modeling techniques are available which allow the incorporation of explanatory variables; for example, Proportional Hazards Modeling, Proportional Intensity Modeling and Additive Hazards Modeling. However, in many applied settings it is unclear what the form of the underlying process is, and thus which of the above modeling structures is the most appropriate, if any. In this paper we discuss the different modeling formulations with regard to such features as their appropriateness, flexibility, robustness and ease of implementation together with the author’s experience gained from application of the models to a wide selection of reliability data sets. In particular, a comparative study of the models when applied to a software reliability data set is provided.

scholarly journals Use of artificial intelligence to estimate population health indicators in France

2020 ◽  
Vol 30 (Supplement_5) ◽  
Author(s):  
R Haneef ◽  
S Fuentes ◽  
R Hrzic ◽  
S Fosse-Edorh ◽  
S Kab ◽  
...  
Keyword(s):  
Artificial Intelligence ◽  
Machine Learning ◽  
Training Data ◽  
Supervised Machine Learning ◽  
Data Sets ◽  
Data Set ◽  
Selection Of

Abstract Background The use of artificial intelligence is increasing to estimate and predict health outcomes from large data sets. The main objectives were to develop two algorithms using machine learning techniques to identify new cases of diabetes (case study I) and to classify type 1 and type 2 (case study II) in France. Methods We selected the training data set from a cohort study linked with French national Health database (i.e., SNDS). Two final datasets were used to achieve each objective. A supervised machine learning method including eight following steps was developed: the selection of the data set, case definition, coding and standardization of variables, split data into training and test data sets, variable selection, training, validation and selection of the model. We planned to apply the trained models on the SNDS to estimate the incidence of diabetes and the prevalence of type 1/2 diabetes. Results For the case study I, 23/3468 and for case study II, 14/3481 SNDS variables were selected based on an optimal balance between variance explained and using the ReliefExp algorithm. We trained four models using different classification algorithms on the training data set. The Linear Discriminant Analysis model performed best in both case studies. The models were assessed on the test datasets and achieved a specificity of 67% and a sensitivity of 62% in case study I, and a specificity of 97 % and sensitivity of 100% in case study II. The case study II model was applied to the SNDS and estimated the prevalence of type 1 diabetes in 2016 in France of 0.3% and for type 2, 4.4%. The case study model I was not applied to the SNDS. Conclusions The case study II model to estimate the prevalence of type 1/2 diabetes has good performance and will be used in routine surveillance. The case study I model to identify new cases of diabetes showed a poor performance due to missing necessary information on determinants of diabetes and will need to be improved for further research.

scholarly journals Extrapolation of Survival Curves Using Standard Parametric Models and Flexible Parametric Spline Models: Comparisons in Large Registry Cohorts with Advanced Cancer

2020 ◽  
pp. 0272989X2097895
Author(s):  
Jodi Gray ◽  
Thomas Sullivan ◽  
Nicholas R. Latimer ◽  
Amy Salter ◽  
Michael J. Sorich ◽  
...  
Keyword(s):  
Survival Data ◽  
Proportional Hazards ◽  
Small Cell ◽  
Parametric Models ◽  
Data Sets ◽  
Small Cell Lung ◽  
Data Set ◽  
Right Censored Data ◽  
Spline Models

Background It is often important to extrapolate survival estimates beyond the limited follow-up times of clinical trials. Extrapolated survival estimates can be highly sensitive to model choice; thus, appropriate model selection is crucial. Flexible parametric spline models have been suggested as an alternative to standard parametric models; however, their ability to extrapolate is not well understood. Aim To determine how well standard parametric and flexible parametric spline models predict survival when fitted to registry cohorts with artificially right-censored follow-up times. Methods Adults with advanced breast, colorectal, small cell lung, non–small cell lung, or pancreatic cancer with a potential follow-up time of 10 y were selected from the SEER 1973–2015 registry data set. Patients were classified into 15 cohorts by cancer and age group at diagnosis (18–59, 60–69, 70+ y). Follow-up times for each cohort were right censored at 20%, 35%, and 50% survival. Standard parametric models (exponential, Weibull, Gompertz, log-logistic, log-normal, generalized gamma) and spline models (proportional hazards, proportional odds, normal/probit) were fitted to the 10-y data set and the 3 right-censored data sets. Predicted 10-y restricted mean survival time and percentage surviving at 10 y were compared with the observed values. Results Across all data sets, the spline odds and spline normal models most frequently gave accurate predictions of 10-y survival outcomes. Visually, spline models tended to demonstrate better fit to the observed hazard functions than standard parametric models, both in the censored and 10-y data. Conclusions In these cohorts, where there was little uncertainty in the observed data, the spline models performed well when extrapolating beyond the observed data. Spline models should be routinely included in the set of models that are fitted when extrapolating cancer survival data.

Association of consensus molecular subtypes (CMS) with time to progression (TTP), progression free survival (PFS), and overall survival (OS) with second-line FOLFIRI ± regorafenib in metastatic colorectal cancer (mCRC).

2019 ◽  
Vol 37 (4_suppl) ◽  
pp. 597-597
Author(s):  
Michael Sangmin Lee ◽  
Sara R. Selitsky ◽  
Joel S. Parker ◽  
James Todd Auman ◽  
Yunhan Wu ◽  
...  
Keyword(s):  
Gene Expression ◽  
Proportional Hazards ◽  
Cox Proportional Hazards ◽  
Training Data ◽  
Prognostic Impact ◽  
Line Treatment ◽  
Second Line ◽  
Data Set ◽  
Significant Difference ◽  
The Impact

597 Background: LCCC1029 was a 2:1 randomized phase II trial of second-line FOLFIRI plus either regorafenib or placebo in mCRC that showed no statistically significant difference in PFS or OS. CMS, defined using gene expression, is prognostic for PFS and OS in previously untreated mCRC, but the impact of CMS in second-line treatment is unclear, as well as its impact on regorafenib efficacy. Methods: RNAseq on archival tumor tissue was successfully performed in 68 LCCC1029 patients (49 on regorafenib, 19 on placebo). A multinomial elastic net CMS classifier was trained using 6 CRC gene expression data sets with known CMS classification. We built our model with only CMS1-4 classified samples and then applied it to normalized and median adjusted RNASeq from LCCC1029 to classify all samples into CMS1-4. TTP, PFS, and OS were compared using Kaplan-Meier method and log-rank tests, and hazard ratios were estimated using Cox proportional hazards method. Results: Our model had > 93% sensitivity and specificity for CMS1-4 in the training data set; the 17% of non-consensus samples in the training data were predominantly labeled CMS2. We classified the LCCC1029 samples as CMS1 (12%), CMS2 (63%), CMS3 (4%), and CMS4 (21%). CMS was prognostic for TTP (log-rank p=0.03), with median for CMS1 of 2.0 months (95% CI 0.0-4.8) versus 5.6 months (5.3-5.9) for CMS2 and 7.8 months (5.5-10.1) for CMS4. There was a trend toward association between CMS and either PFS (log-rank p = 0.11) or OS (log-rank p = 0.085). CMS2 had superior OS compared to CMS1 (HR 0.39, 95% CI 0.17-0.87, p = 0.02). With our limited sample size, we found no significant interaction between CMS and treatment arm for TTP, PFS, or OS. Conclusions: CMS is associated with significant differences in TTP in second-line treatment of mCRC in LCCC1029, and specific CMS types also have differences in OS. Thus, the prognostic impact of CMS extends to second-line treatment in mCRC, meriting further study of CMS classification in additional non-first-line studies.

scholarly journals Radiosensitivity Index (RSI) is not fit to be used for dose-adjustments: a pan-cancer analysis

2021 ◽  
Author(s):  
Hitesh Mistry
Keyword(s):  
Dose Response ◽  
Data Sets ◽  
Survival Times ◽  
Data Set ◽  
Cancer Data ◽  
Wide Range ◽  
Vitro Analysis ◽  
Pan Cancer ◽  
In Vitro Analysis

Radiotherapy has been striving to find markers of radiotherapy sensitivity for decades. In recent years the community has spent significant resources on exploring the wide range of omics data-sets to find that elusive perfect biomarker. One such candidate termed the Radiosensitivity Index, RSI for short, has been heavily publicized as a marker suitable for making dose-adjustments in the clinical setting. However, none of the analyses conducted, thus far, has assessed whether RSI explains enough of the outcome variance to elucidate a dose-response empirically. Here we re-analyze a pan-cancer data-set and find that RSI is no better than random chance at explaining outcome variance, overall survival times. For completeness, we then assessed whether RSI captured a sufficient amount of outcome variance to elucidate a dose-response, it did not. These results suggest that like the initial in-vitro analysis 12 years previously RSI is not a marker of radiotherapy sensitivity and is thus not fit to be used in any dose-adjustment algorithms.

Age at Exposure to Parental Suicide and the Subsequent Risk of Suicide in Young People

Crisis ◽  
2018 ◽  
Vol 39 (1) ◽  
pp. 27-36 ◽  
Author(s):  
Kuan-Ying Lee ◽  
Chung-Yi Li ◽  
Kun-Chia Chang ◽  
Tsung-Hsueh Lu ◽  
Ying-Yeh Chen
Keyword(s):  
Young People ◽  
Proportional Hazards ◽  
Proportional Hazards Model ◽  
Cox Proportional Hazards ◽  
Cox Proportional Hazards Model ◽  
Birth Registry ◽  
Data Set ◽  
Parental Suicide ◽  
The Impact ◽  
Age At Exposure

Abstract. Background: We investigated the age at exposure to parental suicide and the risk of subsequent suicide completion in young people. The impact of parental and offspring sex was also examined. Method: Using a cohort study design, we linked Taiwan's Birth Registry (1978–1997) with Taiwan's Death Registry (1985–2009) and identified 40,249 children who had experienced maternal suicide (n = 14,431), paternal suicide (n = 26,887), or the suicide of both parents (n = 281). Each exposed child was matched to 10 children of the same sex and birth year whose parents were still alive. This yielded a total of 398,081 children for our non-exposed cohort. A Cox proportional hazards model was used to compare the suicide risk of the exposed and non-exposed groups. Results: Compared with the non-exposed group, offspring who were exposed to parental suicide were 3.91 times (95% confidence interval [CI] = 3.10–4.92 more likely to die by suicide after adjusting for baseline characteristics. The risk of suicide seemed to be lower in older male offspring (HR = 3.94, 95% CI = 2.57–6.06), but higher in older female offspring (HR = 5.30, 95% CI = 3.05–9.22). Stratified analyses based on parental sex revealed similar patterns as the combined analysis. Limitations: As only register-­based data were used, we were not able to explore the impact of variables not contained in the data set, such as the role of mental illness. Conclusion: Our findings suggest a prominent elevation in the risk of suicide among offspring who lost their parents to suicide. The risk elevation differed according to the sex of the afflicted offspring as well as to their age at exposure.

Comparative Analysis of Machine Learning Techniques Using Predictive Modeling

2020 ◽  
Vol 13 ◽  
Author(s):  
Ritu Khandelwal ◽  
Hemlata Goyal ◽  
Rajveer Singh Shekhawat
Keyword(s):  
Machine Learning ◽  
Comparative Analysis ◽  
Data Science ◽  
Training Data ◽  
Machine Learning Techniques ◽  
Future Trends ◽  
Data Set ◽  
Learning Stage ◽  
Learning Techniques ◽  
Different Types

Introduction: Machine learning is an intelligent technology that works as a bridge between businesses and data science. With the involvement of data science, the business goal focuses on findings to get valuable insights on available data. The large part of Indian Cinema is Bollywood which is a multi-million dollar industry. This paper attempts to predict whether the upcoming Bollywood Movie would be Blockbuster, Superhit, Hit, Average or Flop. For this Machine Learning techniques (classification and prediction) will be applied. To make classifier or prediction model first step is the learning stage in which we need to give the training data set to train the model by applying some technique or algorithm and after that different rules are generated which helps to make a model and predict future trends in different types of organizations. Methods: All the techniques related to classification and Prediction such as Support Vector Machine(SVM), Random Forest, Decision Tree, Naïve Bayes, Logistic Regression, Adaboost, and KNN will be applied and try to find out efficient and effective results. All these functionalities can be applied with GUI Based workflows available with various categories such as data, Visualize, Model, and Evaluate. Result: To make classifier or prediction model first step is learning stage in which we need to give the training data set to train the model by applying some technique or algorithm and after that different rules are generated which helps to make a model and predict future trends in different types of organizations Conclusion: This paper focuses on Comparative Analysis that would be performed based on different parameters such as Accuracy, Confusion Matrix to identify the best possible model for predicting the movie Success. By using Advertisement Propaganda, they can plan for the best time to release the movie according to the predicted success rate to gain higher benefits. Discussion: Data Mining is the process of discovering different patterns from large data sets and from that various relationships are also discovered to solve various problems that come in business and helps to predict the forthcoming trends. This Prediction can help Production Houses for Advertisement Propaganda and also they can plan their costs and by assuring these factors they can make the movie more profitable.

scholarly journals Building Damage Detection from Post-Event Aerial Imagery Using Single Shot Multibox Detector

2019 ◽  
Vol 9 (6) ◽  
pp. 1128 ◽  
Author(s):  
Yundong Li ◽  
Wei Hu ◽  
Han Dong ◽  
Xueyan Zhang
Keyword(s):  
Machine Learning ◽  
Data Augmentation ◽  
Hurricane Sandy ◽  
Training Data ◽  
Aerial Images ◽  
Detection Methods ◽  
Single Shot ◽  
Data Set ◽  
Augmentation Strategies ◽  
Post Disaster

Using aerial cameras, satellite remote sensing or unmanned aerial vehicles (UAV) equipped with cameras can facilitate search and rescue tasks after disasters. The traditional manual interpretation of huge aerial images is inefficient and could be replaced by machine learning-based methods combined with image processing techniques. Given the development of machine learning, researchers find that convolutional neural networks can effectively extract features from images. Some target detection methods based on deep learning, such as the single-shot multibox detector (SSD) algorithm, can achieve better results than traditional methods. However, the impressive performance of machine learning-based methods results from the numerous labeled samples. Given the complexity of post-disaster scenarios, obtaining many samples in the aftermath of disasters is difficult. To address this issue, a damaged building assessment method using SSD with pretraining and data augmentation is proposed in the current study and highlights the following aspects. (1) Objects can be detected and classified into undamaged buildings, damaged buildings, and ruins. (2) A convolution auto-encoder (CAE) that consists of VGG16 is constructed and trained using unlabeled post-disaster images. As a transfer learning strategy, the weights of the SSD model are initialized using the weights of the CAE counterpart. (3) Data augmentation strategies, such as image mirroring, rotation, Gaussian blur, and Gaussian noise processing, are utilized to augment the training data set. As a case study, aerial images of Hurricane Sandy in 2012 were maximized to validate the proposed method’s effectiveness. Experiments show that the pretraining strategy can improve of 10% in terms of overall accuracy compared with the SSD trained from scratch. These experiments also demonstrate that using data augmentation strategies can improve mAP and mF1 by 72% and 20%, respectively. Finally, the experiment is further verified by another dataset of Hurricane Irma, and it is concluded that the paper method is feasible.

Sign in / Sign up

Export Citation Format

Share Document