mlr3proba: an R package for machine learning in survival analysis.

SUMMARY: As machine learning has become increasingly popular over the last few decades, so too has the number of machine-learning interfaces for implementing these models. Whilst many R libraries exist for machine learning, very few offer extended support for survival analysis. This is problematic considering its importance in fields like medicine, bioinformatics, economics, engineering and more. mlr3proba provides a comprehensive machine-learning interface for survival analysis and connects with mlr3's general model tuning and benchmarking facilities to provide a systematic infrastructure for survival modelling and evaluation. AVAILABILITY AND IMPLEMENTATION: mlr3proba is available under an LGPL-3 licence on CRAN and at https://github.com/mlr-org/mlr3proba, with further documentation at https://mlr3book.mlr-org.com/survival.html.

The Trail Making test: a study of its ability to predict falls in the acute neurological in-patient population.

OBJECTIVE: To determine whether tests of cognitive function and patient-reported outcome measures of motor function can be used to create a machine learning-based predictive tool for falls. DESIGN: Prospective cohort study. SETTING: Tertiary neurological and neurosurgical center. SUBJECTS: In all, 337 in-patients receiving neurosurgical, neurological, or neurorehabilitation-based care. MAIN MEASURES: Binary (Y/N) for falling during the in-patient episode, the Trail Making Test (a measure of attention and executive function) and the Walk-12 (a patient-reported measure of physical function). RESULTS: The principal outcome was a fall during the in-patient stay ( n = 54). The Trail test was identified as the best predictor of falls. Moreover, addition of other variables, did not improve the prediction (Wilcoxon signed-rank P < 0.001). Classical linear statistical modeling methods were then compared with more recent machine learning based strategies, for example, random forests, neural networks, support vector machines. The random forest was the best modeling strategy when utilizing just the Trail Making Test data (Wilcoxon signed-rank P < 0.001) with 68% (± 7.7) sensitivity, and 90% (± 2.3) specificity. CONCLUSION: This study identifies a simple yet powerful machine learning (Random Forest) based predictive model for an in-patient neurological population, utilizing a single neuropsychological test of cognitive function, the Trail Making test.

Problematic internet use as an age-related multifaceted problem: Evidence from a two-site survey.

BACKGROUND AND AIMS: Problematic internet use (PIU; otherwise known as Internet Addiction) is a growing problem in modern societies. There is scarce knowledge of the demographic variables and specific internet activities associated with PIU and a limited understanding of how PIU should be conceptualized. Our aim was to identify specific internet activities associated with PIU and explore the moderating role of age and gender in those associations. METHODS: We recruited 1749 participants aged 18 and above via media advertisements in an Internet-based survey at two sites, one in the US, and one in South Africa; we utilized Lasso regression for the analysis. RESULTS: Specific internet activities were associated with higher problematic internet use scores, including general surfing (lasso ß: 2.1), internet gaming (ß: 0.6), online shopping (ß: 1.4), use of online auction websites (ß: 0.027), social networking (ß: 0.46) and use of online pornography (ß: 1.0). Age moderated the relationship between PIU and role-playing-games (ß: 0.33), online gambling (ß: 0.15), use of auction websites (ß: 0.35) and streaming media (ß: 0.35), with older age associated with higher levels of PIU. There was inconclusive evidence for gender and gender × internet activities being associated with problematic internet use scores. Attention-deficit hyperactivity disorder (ADHD) and social anxiety disorder were associated with high PIU scores in young participants (age ≤ 25, ß: 0.35 and 0.65 respectively), whereas generalized anxiety disorder (GAD) and obsessive-compulsive disorder (OCD) were associated with high PIU scores in the older participants (age > 55, ß: 6.4 and 4.3 respectively). CONCLUSIONS: Many types of online behavior (e.g. shopping, pornography, general surfing) bear a stronger relationship with maladaptive use of the internet than gaming supporting the diagnostic classification of problematic internet use as a multifaceted disorder. Furthermore, internet activities and psychiatric diagnoses associated with problematic internet use vary with age, with public health implications.

Problematic internet use (PIU): Associations with the impulsive-compulsive spectrum. An application of machine learning in psychiatry.

Problematic internet use is common, functionally impairing, and in need of further study. Its relationship with obsessive-compulsive and impulsive disorders is unclear. Our objective was to evaluate whether problematic internet use can be predicted from recognised forms of impulsive and compulsive traits and symptomatology. We recruited volunteers aged 18 and older using media advertisements at two sites (Chicago USA, and Stellenbosch, South Africa) to complete an extensive online survey. State-of-the-art out-of-sample evaluation of machine learning predictive models was used, which included Logistic Regression, Random Forests and Naïve Bayes. Problematic internet use was identified using the Internet Addiction Test (IAT). 2006 complete cases were analysed, of whom 181 (9.0%) had moderate/severe problematic internet use. Using Logistic Regression and Naïve Bayes we produced a classification prediction with a receiver operating characteristic area under the curve (ROC-AUC) of 0.83 (SD 0.03) whereas using a Random Forests algorithm the prediction ROC-AUC was 0.84 (SD 0.03) [all three models superior to baseline models p < 0.0001]. The models showed robust transfer between the study sites in all validation sets [p < 0.0001]. Prediction of problematic internet use was possible using specific measures of impulsivity and compulsivity in a population of volunteers. Moreover, this study offers proof-of-concept in support of using machine learning in psychiatry to demonstrate replicability of results across geographically and culturally distinct settings.

Prediction and Quantification of Individual Athletic Performance of Runners.

We present a novel, quantitative view on the human athletic performance of individual runners. We obtain a predictor for running performance, a parsimonious model and a training state summary consisting of three numbers by application of modern validation techniques and recent advances in machine learning to the thepowerof10 database of British runners' performances (164,746 individuals, 1,417,432 performances). Our predictor achieves an average prediction error (out-of-sample) of e.g. 3.6 min on elite Marathon performances and 0.3 seconds on 100 metres performances, and a lower error than the state-of-the-art in performance prediction (30% improvement, RMSE) over a range of distances. We are also the first to report on a systematic comparison of predictors for running performance. Our model has three parameters per runner, and three components which are the same for all runners. The first component of the model corresponds to a power law with exponent dependent on the runner which achieves a better goodness-of-fit than known power laws in the study of running. Many documented phenomena in quantitative sports science, such as the form of scoring tables, the success of existing prediction methods including Riegel's formula, the Purdy points scheme, the power law for world records performances and the broken power law for world record speeds may be explained on the basis of our findings in a unified way. We provide strong evidence that the three parameters per runner are related to physiological and behavioural parameters, such as training state, event specialization and age, which allows us to derive novel physiological hypotheses relating to athletic performance. We conjecture on this basis that our findings will be vital in exercise physiology, race planning, the study of aging and training regime design.

Prospective evaluation of allogeneic hematopoietic stem-cell transplantation from matched related and matched unrelated donors in younger adults with high-risk acute myeloid leukemia: German-Austrian trial AMLHD98A.

PURPOSE: To assess the impact of allogeneic hematopoietic stem-cell transplantation (HSCT) from matched related donors (MRDs) and matched unrelated donors (MUDs) on outcome in high-risk patients with acute myeloid leukemia (AML) within a prospective multicenter treatment trial. PATIENTS AND METHODS: Between 1998 and 2004, 844 patients (median age, 48 years; range, 16 to 62 years) with AML were enrolled onto protocol AMLHD98A that included a risk-adapted treatment strategy. High risk was defined by the presence of unfavorable cytogenetics and/or by no response to induction therapy. RESULTS: Two hundred sixty-seven (32%) of 844 patients were assigned to the high-risk group. Of these 267 patients, 51 patients (19%) achieved complete remission but had adverse cytogenetics, and 216 patients (81%) had no response to induction therapy. Allogeneic HSCT was actually performed in 162 (61%) of 267 high-risk patients, after a median time of 147 days after diagnosis. Graft sources were as follows: MRD (n = 62), MUD (n = 89), haploidentical donor (n = 10), and cord blood (n = 1). The 5-year overall survival rates were 6.5% (95% CI, 3.1% to 13.6%) for patients (n = 105) not proceeding to HSCT and 25.1% (95% CI, 19.1% to 33.0%; from date of transplantation) for patients (n = 162) receiving HSCT. Multivariable analysis including allogeneic HSCT as a time-dependent covariable revealed that allogeneic HSCT significantly improved outcome; there was no difference in outcome between allogeneic HSCT from MRD and MUD. CONCLUSION: Allogeneic HSCT in younger adults with high-risk AML has a significant beneficial impact on outcome, and allogeneic HSCT from MRD and MUD yields similar results.

Modelling radioimmunotherapy with anti-CD45 antibody to obtain a more favourable biodistribution.

UNLABELLED: Radioimmunotherapy (RIT) is a method to selectively deliver radiation to malignant haematological cells by addressing specific antigens. One approach to improve the biodistribution is to administer a preload of unlabelled antibodies. The aim of this study was to develop a model, which describes distribution of labelled and unlabelled antibodies based on the tissue blood flow and the competing binding behaviour of the antibodies. Such a model can be used to improve biodistribution in the particular case of RIT using anti-CD45 antibodies. METHODS: A compartmental model for the interconnected organs was developed. Reaction constants and organ specific flow, antigen concentrations and distribution volumes were taken from the literature. The organ residence times were calculated for different amounts of given labelled and unlabelled antibodies and the time delay between their administrations. RESULTS: The model is capable to describe the preloading effect. The biodistribution of labelled or unlabelled antibodies depends essentially on the specific blood flow to the organ and its antigen expression. The dose ratio of bone marrow to liver is maximized by applying sufficient unlabelled monoclonal antibody (mAb) to saturate antibody binding in the competing organs and by applying the labelled mAb with a delay of more than one hour. CONCLUSIONS: The developed model qualitatively describes how a preload can considerably increase selectivity of RIT due to different blood flows and antigen distribution in relevant organs. In addition, simulations can identify the optimal delay between the application of labelled and unlabelled antibody. For future analyses, i.e., to fit patient data, degradation and excretion should be incorporated into the model.

Finding stationary subspaces in multivariate time series.

Identifying temporally invariant components in complex multivariate time series is key to understanding the underlying dynamical system and predict its future behavior. In this Letter, we propose a novel technique, stationary subspace analysis (SSA), that decomposes a multivariate time series into its stationary and nonstationary part. The method is based on two assumptions: (a) the observed signals are linear superpositions of stationary and nonstationary sources; and (b) the nonstationarity is measurable in the first two moments. We characterize theoretical and practical properties of SSA and study it in simulations and cortical signals measured by electroencephalography. Here, SSA succeeds in finding stationary components that lead to a significantly improved prediction accuracy and meaningful topographic maps which contribute to a better understanding of the underlying nonstationary brain processes.