A subject being followed over time may experience several types of events related, for example, to disease morbidity and mortality. For example, in a Phase III trial of concomitant versus sequential chemotherapy and thoracic radiotherapy for patients with inoperable non-small cell lung cancer (NSCLC) conducted by the Radiation Therapy Oncology Group (RTOG), patients were followed up to 5 years, the occurrence of either disease progression or death being of particular interest. Such "competing risks" data are commonly encountered in cancer and other biomedical followup studies, in addition to the potential complication of right-censoring on the event time(s) of interest. Two quantities are often used when analyzing competing risks data: the cause-specific hazard function (CSH) and the cumulative incidence function (CIF). For a given event, the former describes the instantaneous risk of this event at time t, given that no events have yet occurred; the latter describes the probability of occurrence, or absolute risk, of that event across time and can be derived directly from the subdistribution hazard function (Fine and Gray, 1999).

Black-box nature hinders the deployment of many high-accuracy models in medical diagnosis. It is risky to put one's life in the hands of models that medical researchers do not trust. However, to understand the mechanism of a new virus, such as COVID-19, machine learning models may catch important symptoms that medical practitioners do not notice due to the surge of infected patients during a pandemic. In this work, the interpretation of machine learning models reveals that a high C-reactive protein (CRP) corresponds to severe infection, and severe patients usually go through a cardiac injury, which is consistent with well-established medical knowledge. Additionally, through the interpretation of machine learning models, we find phlegm and diarrhea are two important symptoms, without which indicate a high risk of turning severe. These two symptoms are not recognized at the early stage of the outbreak, whereas our findings are corroborated by later autopsies of COVID-19 patients. We find patients with a high N-terminal pro B-type natriuretic peptide (NTproBNP) have a significantly increased risk of death which does not receive much attention initially but proves true by the following-up study. Thus, we suggest interpreting machine learning models can offer help to diagnosis at the early stage of an outbreak.

An algorithm for non-stationary spatial modelling using multiple secondary variables is developed. It combines Geostatistics with Quantile Random Forests to give a new interpolation and stochastic simulation algorithm. This paper introduces the method and shows that it has consistency results that are similar in nature to those applying to geostatistical modelling and to Quantile Random Forests. The method allows for embedding of simpler interpolation techniques, such as Kriging, to further condition the model. The algorithm works by estimating a conditional distribution for the target variable at each target location. The family of such distributions is called the envelope of the target variable. From this, it is possible to obtain spatial estimates, quantiles and uncertainty. An algorithm to produce conditional simulations from the envelope is also developed. As they sample from the envelope, realizations are therefore locally influenced by relative changes of importance of secondary variables, trends and variability.

I develop Macroeconomic Random Forest (MRF), an algorithm adapting the canonical Machine Learning (ML) tool to flexibly model evolving parameters in a linear macro equation. Its main output, Generalized Time-Varying Parameters (GTVPs), is a versatile device nesting many popular nonlinearities (threshold/switching, smooth transition, structural breaks/change) and allowing for sophisticated new ones. The approach delivers clear forecasting gains over numerous alternatives, predicts the 2008 drastic rise in unemployment, and performs well for inflation. Unlike most ML-based methods, MRF is directly interpretable -- via its GTVPs. For instance, the successful unemployment forecast is due to the influence of forward-looking variables (e.g., term spreads, housing starts) nearly doubling before every recession. Interestingly, the Phillips curve has indeed flattened, and its might is highly cyclical.

We propose Enhash, a fast ensemble learner that detects \textit{concept drift} in a data stream. A stream may consist of abrupt, gradual, virtual, or recurring events, or a mixture of various types of drift. Enhash employs projection hash to insert an incoming sample. We show empirically that the proposed method has competitive performance to existing ensemble learners in much lesser time. Also, Enhash has moderate resource requirements. Experiments relevant to performance comparison were performed on 6 artificial and 4 real data sets consisting of various types of drifts.

Decision trees and their ensembles are endowed with a rich set of diagnostic tools for ranking and screening input variables in a predictive model. One of the most commonly used in practice is the Mean Decrease in Impurity (MDI), which calculates an importance score for a variable by summing the weighted impurity reductions over all non-terminal nodes split with that variable. Despite the widespread use of tree based variable importance measures such as MDI, pinning down their theoretical properties has been challenging and therefore largely unexplored. To address this gap between theory and practice, we derive rigorous finite sample performance guarantees for variable ranking and selection in nonparametric models with MDI for a single-level CART decision tree (decision stump). We find that the marginal signal strength of each variable and ambient dimensionality can be considerably weaker and higher, respectively, than state-of-the-art nonparametric variable selection methods. Furthermore, unlike previous marginal screening methods that attempt to directly estimate each marginal projection via a truncated basis expansion, the fitted model used here is a simple, parsimonious decision stump, thereby eliminating the need for tuning the number of basis terms. Thus, surprisingly, even though decision stumps are highly inaccurate for estimation purposes, they can still be used to perform consistent model selection.

Predictive clustering trees (PCTs) are a well established generalization of standard decision trees, which can be used to solve a variety of predictive modeling tasks, including structured output prediction. Combining them into ensembles yields state-of-the-art performance. Furthermore, the ensembles of PCTs can be interpreted by calculating feature importance scores from the learned models. However, their learning time scales poorly with the dimensionality of the output space. This is often problematic, especially in (hierarchical) multi-label classification, where the output can consist of hundreds of potential labels. Also, learning of PCTs can not exploit the sparsity of data to improve the computational efficiency, which is common in both input (molecular fingerprints, bag of words representations) and output spaces (in multi-label classification, examples are often labeled with only a fraction of possible labels). In this paper, we propose oblique predictive clustering trees, capable of addressing these limitations. We design and implement two methods for learning oblique splits that contain linear combinations of features in the tests, hence a split corresponds to an arbitrary hyperplane in the input space. The methods are efficient for high dimensional data and capable of exploiting sparse data. We experimentally evaluate the proposed methods on 60 benchmark datasets for 6 predictive modeling tasks. The results of the experiments show that oblique predictive clustering trees achieve performance on-par with state-of-the-art methods and are orders of magnitude faster than standard PCTs. We also show that meaningful feature importance scores can be extracted from the models learned with the proposed methods.

I use one of those credit monitoring services that regularly emails me about my credit score: "Congratulations, your score has gone up!" "Uh oh, your score has gone down!" I shrug and delete the emails. Credit scores are just one example of the many automated decisions made about us as individuals on the basis of complex models. I don't know exactly what causes those little changes in my score. Some machine learning models are "black boxes," a term often used to describe models whose inner workings -- the ways different variables ended up related to one another by an algorithm -- may be impossible for even their designers to completely interpret and explain.

Exploring bias in AI systems, and what we can do to prevent it. For business and non-profit leaders trying to understand AI, it can be surprisingly difficult to find good information in the sweet spot between high-level overview and technical jargon. The AI Clarified series attempts to fill this void and answer some of the most commonly asked AI questions with practical, easy-to-follow explanations. Question: Is AI more biased than humans, or less? I've heard both and am not sure which side to believe. Indeed it's hard to know what to believe about bias in Artificial Intelligence (AI) systems when just reading articles online -- there is plenty of support in both directions. With the growth of AI and the widespread adaption of AI models, there is a lot of noise on both sides, especially for high-stakes use cases like those affecting humans. Let's take hiring as an example.

Survival data with time-varying covariates are common in practice. However, the traditional survival forests - conditional inference forest, relative risk forest and random survival forest - have accommodated only time-invariant covariates. Similarly, the recently proposed transformation forest, which incorporates the split statistics suitable for non-proportional hazard settings, has employed only time-invariant covariates. We generalize the conditional inference and relative risk forests to allow time-varying covariates. We compare their performance with that of the Cox model and transformation forest, adapted to accommodate time-varying covariates, through a comprehensive simulation study in which the Kaplan-Meier estimate serves as a benchmark. In general, the performance of the two proposed forests substantially improves over the Kaplan-Meier estimate when the estimation conditions become more favorable. Taking into an account all other factors, under the PH setting, the best method is always one of the two proposed forests, while under the non-PH setting, it is the adapted transformation forest. The K-fold cross-validation can be an effective tool to choose between the methods in practice. Finally, the performance of the proposed forest methods for time-invariant covariate data is broadly similar to that found for time-varying covariate data. We also propose a general framework for estimation of a survival function in the presence of time-varying covariates, which can be applied to any method that uses the counting process (pseudo-subject) approach to handling time-varying covariates. This novel estimate of a single survival function takes multiple survival estimation outputs corresponding to each pseudo-subject, and combines them in a theoretically-justified way to form a proper monotone-decreasing survival function estimate.