A comparison of different modelling techniques in predicting mortality with the Tilburg Frailty Indicator (TFI) (Preprint)

Modelling Techniques ◽

BACKGROUND Background Modern modelling techniques may potentially provide more accurate predictions of dichotomous outcomes than classical techniques. OBJECTIVE Objective We aimed to study the predictive performance of eight modelling techniques to predict mortality by frailty. METHODS Methods We performed a longitudinal study with a 7-year follow-up. The sample consisted of 479 Dutch community-dwelling people >=75 years. Frailty was assessed with the Tilburg Frailty Indicator (TFI), a self-report questionnaire. This questionnaire consisted of eight physical, four psychological and three social frailty components. The municipality of Roosendaal (a city in the Netherlands) provided the mortality dates. We compared modelling techniques such as support vector machine, neural net, random forest, least absolute shrinkage and selection operator and classical techniques such as logistic regression, two 1Bayesian networks and recursive partitioning. The area under the ROC-curve (AUC) indicated the performance of the models. The models were validated using bootstrapping. RESULTS Results We found that the neural net model had the best validated performance (AUC=0.812) followed by the support vector machine model (AUC=0.705). The other models had validated AUCs <0.700. The recursive partitioning model had the lowest validated AUC (0.605). The neural net model had the highest optimism (0.156). The predictor variable ’difficulty in walking’ was important for all models. CONCLUSIONS Conclusions Because of the high optimism of the NN model, we prefer the SVM model for predicting mortality in community-dwelling older people with the TFI with added to it ’gender’ and ’age’. External validation is a necessary step before applying the prediction models in a new setting.

Psychometric Properties of the Taiwanese Version of the Tilburg Frailty Indicator for Community-Dwelling Older Adults

Healthcare ◽

10.3390/healthcare9091193 ◽

2021 ◽

Vol 9 (9) ◽

pp. 1193

Author(s):

Chia-Hui Lin ◽

Chieh-Yu Liu ◽

Jiin-Ru Rong

Keyword(s):

Older Adults ◽

Psychometric Properties ◽

Roc Curve ◽

Goodness Of Fit ◽

Community Dwelling ◽

Self Report ◽

Youden Index ◽

Community Dwelling Older Adults ◽

Screening the frailty level of older adults is essential to avoid morbidity, prevent falls and disability, and maintain quality of life. The Tilburg Frailty Indicator (TFI) is a self-report instrument developed to assess frailty for community-dwelling older adults. The aim of this study was to explore the psychometric properties of the Taiwanese version of TFI (TFI-T). The sample consisted of 210 elderly participants living in the community. The scale was implemented to conduct a confirmatory factor analysis (CFA) test for validity. The models were evaluated through sensitivity, specificity, area under the curve, and receiving operating characteristic (ROC) curve. CFA was performed to evaluate construct validity, and the TFI-T has a goodness of fit with the three-factor structure of the TFI. Totally, the 15 items of TFI-T have acceptable internal consistency (Cronbach’s alpha = 0.78), and test–retest reliability (r = 0.88, p < 0.001). The criterion-related validity was examined, the TFI-T correlation with the Kihon Checklist (KCL) score (r = 0.74; p < 0.001). The cutoff of 5.5 based on the Youden index was considered optimal. The area under the ROC curve analysis indicated that the TFI-T has good accuracy in frailty screening. The TFI-T exhibits good reliability and validity and can be used as a sensitive and accurate instrument, which is highly applicable to screen frailty in Taiwan among older adults.

Classification and Regression Models for Genomic Selection of Skewed Phenotypes: A Case for Disease Resistance in Winter Wheat (Triticum aestivum L.)

10.1101/2021.12.16.472985 ◽

2021 ◽

Author(s):

Lance F Merrick ◽

Dennis N Lozada ◽

Xianming Chen ◽

Arron H Carter

Keyword(s):

Support Vector Machine ◽

Winter Wheat ◽

Genomic Selection ◽

Stripe Rust ◽

Regression Models ◽

Prediction Models ◽

Support Vector ◽

Classification Models ◽

Breeding Lines ◽

Classification And Regression

Most genomic prediction models are linear regression models that assume continuous and normally distributed phenotypes, but responses to diseases such as stripe rust (caused by Puccinia striiformis f. sp. tritici) are commonly recorded in ordinal scales and percentages. Disease severity (SEV) and infection type (IT) data in germplasm screening nurseries generally do not follow these assumptions. On this regard, researchers may ignore the lack of normality, transform the phenotypes, use generalized linear models, or use supervised learning algorithms and classification models with no restriction on the distribution of response variables, which are less sensitive when modeling ordinal scores. The goal of this research was to compare classification and regression genomic selection models for skewed phenotypes using stripe rust SEV and IT in winter wheat. We extensively compared both regression and classification prediction models using two training populations composed of breeding lines phenotyped in four years (2016-2018, and 2020) and a diversity panel phenotyped in four years (2013-2016). The prediction models used 19,861 genotyping-by-sequencing single-nucleotide polymorphism markers. Overall, square root transformed phenotypes using rrBLUP and support vector machine regression models displayed the highest combination of accuracy and relative efficiency across the regression and classification models. Further, a classification system based on support vector machine and ordinal Bayesian models with a 2-Class scale for SEV reached the highest class accuracy of 0.99. This study showed that breeders can use linear and non-parametric regression models within their own breeding lines over combined years to accurately predict skewed phenotypes.

Algorithmic and data modeling: Will algorithmic modeling improve predictions of traits evaluated on ordinal scales?

10.1101/2020.10.07.329466 ◽

2020 ◽

Author(s):

Zhanyou Xu ◽

Andreomar Kurek ◽

Steven B. Cannon ◽

Williams D. Beavis

Keyword(s):

Support Vector Machine ◽

Random Forest ◽

Ridge Regression ◽

Genomic Prediction ◽

Ordinal Data ◽

Prediction Models ◽

Characteristic Curve ◽

Gradient Boosting ◽

Support Vector ◽

Data Types

AbstractSelection of markers linked to alleles at quantitative trait loci (QTL) for tolerance to Iron Deficiency Chlorosis (IDC) has not been successful. Genomic selection has been advocated for continuous numeric traits such as yield and plant height. For ordinal data types such as IDC, genomic prediction models have not been systematically compared. The objectives of research reported in this manuscript were to evaluate the most commonly used genomic prediction method, ridge regression and it’s equivalent logistic ridge regression method, with algorithmic modeling methods including random forest, gradient boosting, support vector machine, K-nearest neighbors, Naïve Bayes, and artificial neural network using the usual comparator metric of prediction accuracy. In addition we compared the methods using metrics of greater importance for decisions about selecting and culling lines for use in variety development and genetic improvement projects. These metrics include specificity, sensitivity, precision, decision accuracy, and area under the receiver operating characteristic curve. We found that Support Vector Machine provided the best specificity for culling IDC susceptible lines, while Random Forest GP models provided the best combined set of decision metrics for retaining IDC tolerant and culling IDC susceptible lines.

Understanding nitrogen transport in the unsaturated zone with fluctuations in groundwater depth

Water Science & Technology Water Supply ◽

10.2166/ws.2021.066 ◽

2021 ◽

Author(s):

Jianmin Bian ◽

Qian Wang ◽

Siyu Nie ◽

Hanli Wan ◽

Juanjuan Wu

Keyword(s):

Support Vector Machine ◽

Unsaturated Zone ◽

Prediction Models ◽

Groundwater Depth ◽

Support Vector ◽

Nitrogen Transport ◽

Irrigation Area ◽

Leaching Loss ◽

The Relationship ◽

Hydrus 1D

Abstract Fluctuations in groundwater depth play an important role and are often overlooked when considering the transport of nitrogen in the unsaturated zone. To evaluate directly the variation of nitrogen transport due to fluctuations in groundwater depth, the prediction model of groundwater depth and nitrogen transport were combined and applied by least squares support vector machine and Hydrus-1D in the western irrigation area of Jilin in China. The calibration and testing results showed the prediction models were reliable. Considering different groundwater depth, the concentration of nitrogen was affected significantly with a groundwater depth of 3.42–1.71 m, while it was not affected with groundwater depth of 5.48–6.47 m. The total leaching loss of nitrogen gradually increased with the continuous decrease of groundwater depth. Furthermore, the limited groundwater depth of 1.7 m was found to reduce the risk of nitrogen pollution. This paper systematically analyzes the relationship between groundwater depth and nitrogen transport to form appropriate agriculture strategies.

Study on Crash Injury Severity Prediction of Autonomous Vehicles for Different Emergency Decisions Based on Support Vector Machine Model

Electronics ◽

10.3390/electronics7120381 ◽

2018 ◽

Vol 7 (12) ◽

pp. 381 ◽

Cited By ~ 6

Author(s):

Yaping Liao ◽

Junyou Zhang ◽

Shufeng Wang ◽

Sixian Li ◽

Jian Han

Keyword(s):

Support Vector Machine ◽

Autonomous Vehicles ◽

Injury Severity ◽

Prediction Models ◽

Motor Vehicle ◽

Support Vector ◽

Severity Prediction ◽

Emergency Decision ◽

The Impact ◽

Crash Injury Severity

Motor vehicle crashes remain a leading cause of life and property loss to society. Autonomous vehicles can mitigate the losses by making appropriate emergency decision, and the crash injury severity prediction model is the basis for autonomous vehicles to make decisions in emergency situations. In this paper, based on the support vector machine (SVM) model and NASS/GES crash data, three SVM crash injury severity prediction models (B-SVM, T-SVM, and BT-SVM) corresponding to braking, turning, and braking + turning respectively are established. The vehicle relative speed (REL_SPEED) and the gross vehicle weight rating (GVWR) are introduced into the impact indicators of the prediction models. Secondly, the ordered logit (OL) and back propagation neural network (BPNN) models are established to validate the accuracy of the SVM models. The results show that the SVM models have the best performance than the other two. Next, the impact of REL_SPEED and GVWR on injury severity is analyzed quantitatively by the sensitivity analysis, the results demonstrate that the increase of REL_SPEED and GVWR will make vehicle crash more serious. Finally, the same crash samples under normal road and environmental conditions are input into B-SVM, T-SVM, and BT-SVM respectively, the output results are compared and analyzed. The results show that with other conditions being the same, as the REL_SPEED increased from the low (0–20 mph) to middle (20–45 mph) and then to the high range (45–75 mph), the best emergency decision with the minimum crash injury severity will gradually transition from braking to turning and then to braking + turning.

Prevalence and factors associated with frailty among community-dwelling older people in rural Thanjavur district of South India: a cross-sectional study

BMJ Open ◽

10.1136/bmjopen-2019-032904 ◽

2019 ◽

Vol 9 (10) ◽

pp. e032904 ◽

Cited By ~ 3

Author(s):

Kirubakaran Kesavan Kendhapedi ◽

Niveditha Devasenapathy

Keyword(s):

Older People ◽

Fear Of Falling ◽

Frailty Index ◽

Cross Sectional Study ◽

Community Dwelling ◽

Cross Sectional ◽

Domain Definition ◽

Community Dwelling Older People ◽

ObjectiveThere is sparse data on the prevalence of frailty from rural parts of India. Our aim was to estimate prevalence of frailty among community-dwelling older people in rural South Indian population and explore socio-demographic factors associated with frailty. We further explored the associations between frailty with fear of falling and falls.DesignCommunity based cross-sectional study.SettingFour villages in Thanjavur district of Southern India.ParticipantsRandom sample of adults aged 60 years and above from four villages.MethodsWe sampled community-dwelling older adults from the electoral list of four villages using stratified random sampling. We report prevalence of frailty as defined by physical definition (Fried’s Phenotype), accumulation of deficits (Frailty Index) and multi-domain definition (Tilburg Frailty Indicator). We report proportion of agreement of frailty status between the frailty tools. We used logistic regressions with robust SEs to examine the associations between socio-demographic determinants with frailty and the association between frailty with fear of falling and falls.ResultsAmong the 408 participants, the weighted (non-response and poststratification for sex) prevalence and 95% CI of frailty was 28% (18.9 to 28.1) for physical definition, 59% (53.9 to 64.3) for accumulation of deficits and 63% (57.4 to 67.6) for multi-domain definition. Frailty Index and Tilburg Frailty Indicator had good agreement (80%). Age, female, lower education, lower socioeconomic status, minimum physical activity in routine work were independently associated with frailty irrespective of the frailty definitions. Frail elderly had higher odds of falls as well as fear of falling compared with non-frail, irrespective of the definitions.ConclusionPrevalence of frailty among older people in rural Thanjavur district of South India was high compared with low-income and middle-income countries. Understanding the modifiable determinants of frailty can provide a valuable reference for future prevention and intervention.

The Prediction of Human Abdominal Adiposity Based on the Combination of a Particle Swarm Algorithm and Support Vector Machine

International Journal of Environmental Research and Public Health ◽

10.3390/ijerph17031117 ◽

2020 ◽

Vol 17 (3) ◽

pp. 1117

Author(s):

Xiue Gao ◽

Wenxue Xie ◽

Shifeng Chen ◽

Junjie Yang ◽

Bo Chen

Keyword(s):

Support Vector Machine ◽

Sample Size ◽

Size Dependence ◽

Prediction Models ◽

Test Group ◽

Support Vector ◽

Particle Swarm Algorithm ◽

Abdominal Adiposity ◽

Modeling Group ◽

Svm Algorithm

Background: Abdominal adiposity is an important risk factor of chronic cardiovascular diseases, thus the prediction of abdominal adiposity and obesity can reduce the risks of contracting such diseases. However, the current prediction models display low accuracy and high sample size dependence. The purpose of this study is to put forward a new prediction method based on an improved support vector machine (SVM) to solve these problems. Methods: A total of 200 individuals participated in this study and were further divided into a modeling group and a test group. Their physiological parameters (height, weight, age, the four parameters of abdominal impedance and body fat mass) were measured using the body composition tester (the universal INBODY measurement device) based on BIA. Intelligent algorithms were used in the modeling group to build predictive models and the test group was used in model performance evaluation. Firstly, the optimal boundary C and parameter gamma were optimized by the particle swarm algorithm. We then developed an algorithm to classify human abdominal adiposity according to the parameter setup of the SVM algorithm and constructed the prediction model using this algorithm. Finally, we designed experiments to compare the performances of the proposed method and the other methods. Results: There are different abdominal obesity prediction models in the 1 KHz and 250 KHz frequency bands. The experimental data demonstrates that for the frequency band of 250 KHz, the proposed method can reduce the false classification rate by 10.7%, 15%, and 33% in relation to the sole SVM algorithm, the regression model, and the waistline measurement model, respectively. For the frequency band of 1 KHz, the proposed model is still more accurate. (4) Conclusions: The proposed method effectively improves the prediction accuracy and reduces the sample size dependence of the algorithm, which can provide a reference for abdominal obesity.

Maximum a posteriori Threshold Genomic Prediction Model for Ordinal Traits

G3 Genes|Genome|Genetics ◽

10.1534/g3.120.401733 ◽

2020 ◽

Vol 10 (11) ◽

pp. 4083-4102

Author(s):

Abelardo Montesinos-López ◽

Humberto Gutierrez-Pulido ◽

Osval Antonio Montesinos-López ◽

José Crossa

Keyword(s):

Support Vector Machine ◽

Prediction Model ◽

Ridge Regression ◽

Genomic Prediction ◽

Prediction Models ◽

Prediction Performance ◽

Support Vector ◽

Successful Implementation ◽

A Posteriori ◽

Ordinal Traits

Due to the ever-increasing data collected in genomic breeding programs, there is a need for genomic prediction models that can deal better with big data. For this reason, here we propose a Maximum a posteriori Threshold Genomic Prediction (MAPT) model for ordinal traits that is more efficient than the conventional Bayesian Threshold Genomic Prediction model for ordinal traits. The MAPT performs the predictions of the Threshold Genomic Prediction model by using the maximum a posteriori estimation of the parameters, that is, the values of the parameters that maximize the joint posterior density. We compared the prediction performance of the proposed MAPT to the conventional Bayesian Threshold Genomic Prediction model, the multinomial Ridge regression and support vector machine on 8 real data sets. We found that the proposed MAPT was competitive with regard to the multinomial and support vector machine models in terms of prediction performance, and slightly better than the conventional Bayesian Threshold Genomic Prediction model. With regard to the implementation time, we found that in general the MAPT and the support vector machine were the best, while the slowest was the multinomial Ridge regression model. However, it is important to point out that the successful implementation of the proposed MAPT model depends on the informative priors used to avoid underestimation of variance components.

Damage Simulation and Ultrasonic Detection of Asphalt Mixture under the Coupling Effects of Water-Temperature-Radiation

Advances in Materials Science and Engineering ◽

10.1155/2013/838943 ◽

2013 ◽

Vol 2013 ◽

pp. 1-9 ◽

Cited By ~ 2

Author(s):

Yong-chun Cheng ◽

Peng Zhang ◽

Yu-bo Jiao ◽

Ye-dan Wang ◽

Jing-lin Tao

Keyword(s):

Support Vector Machine ◽

Water Temperature ◽

Ultrasonic Velocity ◽

Prediction Models ◽

Asphalt Mixture ◽

Support Vector ◽

Ultrasonic Detection ◽

Damage Coefficient ◽

Temperature Radiation ◽

Damage Simulation

In order to accurately simulate the performance changes of asphalt pavement in the hot rainy days, laboratory water-temperature-radiation cycle test is designed and carried out for the damage simulation of asphalt mixture under the environmental effect of rain, high temperatures, and sunshine. Ultrasonic detection method is used to determine the ultrasonic velocity of asphalt mixture specimen under different temperatures and water contents in the process of water-temperature-radiation cycles. Thus, we get the preliminary damage assessment. Splitting strength attenuation is defined as the damage parameter. In addition, the regression prediction models of the ultrasonic velocity and damage coefficient of asphalt mixture are constructed using the grey theory, neural network method, and support vector machine theory, respectively. We compare the prediction results of the three different models. It can be concluded that the model derived from the support vector machine possesses higher accuracy and stability, which can more satisfactorily reflect the relationship between ultrasonic velocity and damage coefficient. Therefore, the damage degree of the asphalt mixture can be obtained.

Comparison of the frailty phenotype and the Tilburg Frailty Indicator regarding the prediction of quality of life in a two-year follow-up

European Psychiatry ◽

10.1016/j.eurpsy.2016.01.405 ◽

2016 ◽

Vol 33 (S1) ◽

pp. S185-S185

Author(s):

T. Coelho ◽

C. Paúl ◽

L. Fernandes

Keyword(s):

Quality Of Life ◽

Adverse Outcomes ◽

Community Dwelling ◽

Stressful Events ◽

Early Screening ◽

Frailty Phenotype ◽

IntroductionFrail individuals are highly vulnerable to minor stressful events, presenting a higher risk for adverse health outcomes (e.g. falls, disability, hospitalization), which can lead to a decline in quality of life (QoL). In this context, an early screening of elderly frailty is of crucial importance.ObjectiveTo compare how the Frailty Phenotype (FP) and the Tilburg Frailty Indicator (TFI) predict QoL in a two-year follow-up.MethodsA longitudinal study was designed recruiting 110 community-dwelling elderly (≥ 65 years). The presence of frailty was assessed at baseline (FP ≥ 3 and TFI ≥ 6), whereas QoL was measured two years later with two different scales: the WHOQOL-OLD and the EUROHIS-QOL-8. Hierarchical regressions were conducted.ResultsThe mean age of the participants at baseline was 77.7 ± 6.9 years, and most were women (75.5%). According to FP, 33.6% of the participants were classified as frail, while the TFI detected frailty in 50% of the elderly. After adjusting for age and gender, the TFI significantly predicted QoL (WHOQOL-OLD: β = −18.9, t(106) = −6.97, P < 0.001; EUROHIS-QOL-8: β = −6.1, t(106) = −6.71, P < 0.001), whereas the effect of the FP on the outcome measures was non-significant.ConclusionsFrailty at baseline was associated with a lower QoL at follow-up. A multidimensional frailty operationalization (TFI) showed a stronger predictive validity than an exclusively physical one (FP). The option of which frailty measure to use in a clinical setting should take into account its ability to predict specific adverse outcomes, conducing to targeted and effective interventions.Disclosure of interestThe authors have not supplied their declaration of competing interest.