arXiv subjects

J. Derek Tucker

Publications and source records attributed to J. Derek Tucker.

16 recordsLinked to original sources

arXiv2025

Bayesian Adaptive Polynomial Chaos Expansions

Polynomial chaos expansions (PCE) are widely used for uncertainty quantification (UQ) tasks, particularly in the applied mathematics community. However, PCE has received comparatively less attention in the statistics literature, and fully Bayesian formulations remain rare, especially with implementations in R. Motivated by the success of adaptive Bayesian machine learning models such as BART, BASS, and BPPR, we develop a new fully Bayesian adaptive PCE method with an efficient and accessible R implementation: khaos. Our approach includes a novel proposal distribution that enables data-driven interaction selection, and supports a modified g-prior tailored to PCE structure. Through simulation studies and real-world UQ applications, we demonstrate that Bayesian adaptive PCE provides competitive performance for surrogate modeling, global sensitivity analysis, and ordinal regression tasks.

stat.ME↗

arXiv2025

Conformal Anomaly Detection for Functional Data with Elastic Distance Metrics

This paper considers the problem of outlier detection in functional data analysis focusing particularly on the more difficult case of shape outliers. We present an inductive conformal anomaly detection method based on elastic functional distance metrics. This method is evaluated and compared to similar conformal anomaly detection methods for functional data using simulation experiments. The method is also used in the analysis of two real exemplar data sets that show its utility in practical applications. The results demonstrate the efficacy of the proposed method for detecting both magnitude and shape outliers in two distinct outlier detection scenarios.

stat.ME↗

arXiv2025

An Explainable Pipeline for Machine Learning with Functional Data

Machine learning (ML) models have shown success in applications with an objective of prediction, but the algorithmic complexity of some models makes them difficult to interpret. Methods have been proposed to provide insight into these "black-box" models, but there is little research that focuses on supervised ML when the model inputs are functional data. In this work, we consider two applications from high-consequence spaces with objectives of making predictions using functional data inputs. One application aims to classify material types to identify explosive materials given hyperspectral computed tomography scans of the materials. The other application considers the forensics science task of connecting an inkjet printed document to the source printer using color signatures extracted by Raman spectroscopy. An instinctive route to consider for analyzing these data is a data driven ML model for classification, but due to the high consequence nature of the applications, we argue it is important to appropriately account for the nature of the data in the analysis to not obscure or misrepresent patterns. As such, we propose the Variable importance Explainable Elastic Shape Analysis (VEESA) pipeline for training ML models with functional data that (1) accounts for the vertical and horizontal variability in the functional data and (2) provides an explanation in the original data space of how the model uses variability in the functional data for prediction. The pipeline makes use of elastic functional principal components analysis (efPCA) to generate uncorrelated model inputs and permutation feature importance (PFI) to identify the principal components important for prediction. The variability captured by the important principal components in visualized the original data space. We ultimately discuss ideas for natural extensions of the VEESA pipeline and challenges for future research.

cs.LG↗

arXiv2024

Elastic Bayesian Model Calibration

Functional data are ubiquitous in scientific modeling. For instance, quantities of interest are modeled as functions of time, space, energy, density, etc. Uncertainty quantification methods for computer models with functional response have resulted in tools for emulation, sensitivity analysis, and calibration that are widely used. However, many of these tools do not perform well when the computer model's parameters control both the amplitude variation of the functional output and its alignment (or phase variation). This paper introduces a framework for Bayesian model calibration when the model responses are misaligned functional data. The approach generates two types of data out of the misaligned functional responses: (1) aligned functions so that the amplitude variation is isolated and (2) warping functions that isolate the phase variation. These two types of data are created for the computer simulation data (both of which may be emulated) and the experimental data. The calibration approach uses both types so that it seeks to match both the amplitude and phase of the experimental data. The framework is careful to respect constraints that arise especially when modeling phase variation, and is framed in a way that it can be done with readily available calibration software. We demonstrate the techniques on two simulated data examples and on two dynamic material science problems: a strength model calibration using flyer plate experiments and an equation of state model calibration using experiments performed on the Sandia National Laboratories' Z-machine.

stat.ME↗

arXiv2023

Detecting changepoints in globally-indexed functional time series

In environmental and climate data, there is often an interest in determining if and when changes occur in a system. Such changes may result from localized sources in space and time like a volcanic eruption or climate geoengineering events. Detecting such events and their subsequent influence on climate has important policy implications. However, the climate system is complex, and such changes can be challenging to detect. One statistical perspective for changepoint detection is functional time series, where one observes an entire function at each time point. We will consider the context where each time point is a year, and we observe a function of temperature indexed by day of the year. Furthermore, such data is measured at many spatial locations on Earth, which motivates accommodating sets of functional time series that are spatially-indexed on a sphere. Simultaneously inferring changes that can occur at different times for different locations is challenging. We propose test statistics for detecting these changepoints, and we evaluate performance using varying levels of data complexity, including a simulation study, simplified climate model simulations, and climate reanalysis data. We evaluate changes in stratospheric temperature globally over 1984-1998. Such changes may be associated with the eruption of Mt. Pinatubo in 1991.

stat.AP↗

arXiv2023

Elastic Functional Changepoint Detection of Climate Impacts from Localized Sources

Detecting changepoints in functional data has become an important problem as interest in monitoring of climate phenomenon has increased, where the data is functional in nature. The observed data often contains both amplitude ($y$-axis) and phase ($x$-axis) variability. If not accounted for properly, true changepoints may be undetected, and the estimated underlying mean change functions will be incorrect. In this paper, an elastic functional changepoint method is developed which properly accounts for these types of variability. The method can detect amplitude and phase changepoints which current methods in the literature do not, as they focus solely on the amplitude changepoint. This method can easily be implemented using the functions directly or can be computed via functional principal component analysis to ease the computational burden. We apply the method and its non-elastic competitors to both simulated data and observed data to show its efficiency in handling data with phase variation with both amplitude and phase changepoints. We use the method to evaluate potential changes in stratospheric temperature due to the eruption of Mt.\ Pinatubo in the Philippines in June 1991. Using an epidemic changepoint model, we find evidence of a increase in stratospheric temperature during a period that contains the immediate aftermath of Mt.\ Pinatubo, with most detected changepoints occurring in the tropics as expected.

stat.ME↗

arXiv2023

Dimensionality Reduction using Elastic Measures

With the recent surge in big data analytics for hyper-dimensional data there is a renewed interest in dimensionality reduction techniques for machine learning applications. In order for these methods to improve performance gains and understanding of the underlying data, a proper metric needs to be identified. This step is often overlooked and metrics are typically chosen without consideration of the underlying geometry of the data. In this paper, we present a method for incorporating elastic metrics into the t-distributed Stochastic Neighbor Embedding (t-SNE) and Uniform Manifold Approximation and Projection (UMAP). We apply our method to functional data, which is uniquely characterized by rotations, parameterization, and scale. If these properties are ignored, they can lead to incorrect analysis and poor classification performance. Through our method we demonstrate improved performance on shape identification tasks for three benchmark data sets (MPEG-7, Car data set, and Plane data set of Thankoor), where we achieve 0.77, 0.95, and 1.00 F1 score, respectively.

cs.LG↗

arXiv2021

Spatio-temporal extreme event modeling of terror insurgencies

Extreme events with potential deadly outcomes, such as those organized by terror groups, are highly unpredictable in nature and an imminent threat to society. In particular, quantifying the likelihood of a terror attack occurring in an arbitrary space-time region and its relative societal risk, would facilitate informed measures that would strengthen national security. This paper introduces a novel self-exciting marked spatio-temporal model for attacks whose inhomogeneous baseline intensity is written as a function of covariates. Its triggering intensity is succinctly modeled with a Gaussian Process prior distribution to flexibly capture intricate spatio-temporal dependencies between an arbitrary attack and previous terror events. By inferring the parameters of this model, we highlight specific space-time areas in which attacks are likely to occur. Furthermore, by measuring the outcome of an attack in terms of the number of casualties it produces, we introduce a novel mixture distribution for the number of casualties. This distribution flexibly handles low and high number of casualties and the discrete nature of the data through a {\it Generalized ZipF} distribution. We rely on a customized Markov chain Monte Carlo (MCMC) method to estimate the model parameters. We illustrate the methodology with data from the open source Global Terrorism Database (GTD) that correspond to attacks in Afghanistan from 2013-2018. We show that our model is able to predict the intensity of future attacks for 2019-2021 while considering various covariates of interest such as population density, number of regional languages spoken, and the density of population supporting the opposing government.

stat.AP↗

arXiv2021

Multimodal Bayesian Registration of Noisy Functions using Hamiltonian Monte Carlo

Functional data registration is a necessary processing step for many applications. The observed data can be inherently noisy, often due to measurement error or natural process uncertainty, which most functional alignment methods cannot handle. A pair of functions can also have multiple optimal alignment solutions, which is not addressed in current literature. In this paper, a flexible Bayesian approach to functional alignment is presented, which appropriately accounts for noise in the data without any pre-smoothing required. Additionally, by running parallel MCMC chains, the method can account for multiple optimal alignments via the multi-modal posterior distribution of the warping functions. To most efficiently sample the warping functions, the approach relies on a modification of the standard Hamiltonian Monte Carlo to be well-defined on the infinite-dimensional Hilbert space. This flexible Bayesian alignment method is applied to both simulated data and real data sets to show its efficiency in handling noisy functions and successfully accounting for multiple optimal alignments in the posterior; characterizing the uncertainty surrounding the warping functions.

stat.ME↗

arXiv2020

Elastic $k$-means clustering of functional data for posterior exploration, with an application to inference on acute respiratory infection dynamics

We propose a new method for clustering of functional data using a $k$-means framework. We work within the elastic functional data analysis framework, which allows for decomposition of the overall variation in functional data into amplitude and phase components. We use the amplitude component to partition functions into shape clusters using an automated approach. To select an appropriate number of clusters, we additionally propose a novel Bayesian Information Criterion defined using a mixture model on principal components estimated using functional Principal Component Analysis. The proposed method is motivated by the problem of posterior exploration, wherein samples obtained from Markov chain Monte Carlo algorithms are naturally represented as functions. We evaluate our approach using a simulated dataset, and apply it to a study of acute respiratory infection dynamics in San Luis Potosí, Mexico.

stat.ME↗

arXiv2020

A Deterministic Hitting-Time Moment Approach to Seed-set Expansion over a Graph

We introduce HITMIX, a new technique for network seed-set expansion, i.e., the problem of identifying a set of graph vertices related to a given seed-set of vertices. We use the moments of the graph's hitting-time distribution to quantify the relationship of each non-seed vertex to the seed-set. This involves a deterministic calculation for the hitting-time moments that is scalable in the number of graph edges and so avoids directly sampling a Markov chain over the graph. The moments are used to fit a mixture model to estimate the probability that each non-seed vertex should be grouped with the seed set. This membership probability enables us to sort the non-seeds and threshold in a statistically-justified way. To the best of our knowledge, HITMIX is the first full statistical model for seed-set expansion that can give vertex-level membership probabilities. While HITMIX is a global method, its linear computation complexity in practice enables computations on large graphs. We have a high-performance implementation, and we present computational results on stochastic blockmodels and a small-world network from the SNAP repository. The state of the art in this problem is a collection of recently developed local methods, and we show that distinct advantages in solution quality are available if our global method can be used. In practice, we expect to be able to run HITMIX if the graph can be stored in memory.

cs.SI↗

arXiv2020

Regression Models Using Shapes of Functions as Predictors

Functional variables are often used as predictors in regression problems. A commonly-used parametric approach, called {\it scalar-on-function regression}, uses the $\ltwo$ inner product to map functional predictors into scalar responses. This method can perform poorly when predictor functions contain undesired phase variability, causing phases to have disproportionately large influence on the response variable. One past solution has been to perform phase-amplitude separation (as a pre-processing step) and then use only the amplitudes in the regression model. Here we propose a more integrated approach, termed elastic functional regression model (EFRM), where phase-separation is performed inside the regression model, rather than as a pre-processing step. This approach generalizes the notion of phase in functional data, and is based on the norm-preserving time warping of predictors. Due to its invariance properties, this representation provides robustness to predictor phase variability and results in improved predictions of the response variable over traditional models. We demonstrate this framework using a number of datasets involving gait signals, NMR data, and stock market prices.

stat.ME↗

arXiv2019

A Geometric Approach for Computing Tolerance Bounds for Elastic Functional Data

We develop a method for constructing tolerance bounds for functional data with random warping variability. In particular, we define a generative, probabilistic model for the amplitude and phase components of such observations, which parsimoniously characterizes variability in the baseline data. Based on the proposed model, we define two different types of tolerance bounds that are able to measure both types of variability, and as a result, identify when the data has gone beyond the bounds of amplitude and/or phase. The first functional tolerance bounds are computed via a bootstrap procedure on the geometric space of amplitude and phase functions. The second functional tolerance bounds utilize functional Principal Component Analysis to construct a tolerance factor. This work is motivated by two main applications: process control and disease monitoring. The problem of statistical analysis and modeling of functional data in process control is important in determining when a production has moved beyond a baseline. Similarly, in biomedical applications, doctors use long, approximately periodic signals (such as the electrocardiogram) to diagnose and monitor diseases. In this context, it is desirable to identify abnormalities in these signals. We additionally consider a simulated example to assess our approach and compare it to two existing methods.

stat.CO↗

arXiv2019

Computer model calibration based on image warping metrics: an application for sea ice deformation

Arctic sea ice plays an important role in the global climate. Sea ice models governed by physical equations have been used to simulate the state of the ice including characteristics such as ice thickness, concentration, and motion. More recent models also attempt to capture features such as fractures or leads in the ice. These simulated features can be partially misaligned or misshapen when compared to observational data, whether due to numerical approximation or incomplete physics. In order to make realistic forecasts and improve understanding of the underlying processes, it is necessary to calibrate the numerical model to field data. Traditional calibration methods based on generalized least-square metrics are flawed for linear features such as sea ice cracks. We develop a statistical emulation and calibration framework that accounts for feature misalignment and misshapenness, which involves optimally aligning model output with observed features using cutting edge image registration techniques. This work can also have application to other physical models which produce coherent structures.

stat.AP↗

arXiv2018

Elastic Functional Principal Component Regression

We study regression using functional predictors in situations where these functions contain both phase and amplitude variability. In other words, the functions are misaligned due to errors in time measurements, and these errors can significantly degrade both model estimation and prediction performance. The current techniques either ignore the phase variability, or handle it via pre-processing, i.e., use an off-the-shelf technique for functional alignment and phase removal. We develop a functional principal component regression model which has comprehensive approach in handling phase and amplitude variability. The model utilizes a mathematical representation of the data known as the square-root slope function. These functions preserve the $\mathbf{L}^2$ norm under warping and are ideally suited for simultaneous estimation of regression and warping parameters. Using both simulated and real-world data sets, we demonstrate our approach and evaluate its prediction performance relative to current models. In addition, we propose an extension to functional logistic and multinomial logistic regression

stat.AP↗

arXiv2012

Generative Models for Functional Data using Phase and Amplitude Separation

Constructing generative models for functional observations is an important task in statistical functional analysis. In general, functional data contains both phase (or x or horizontal) and amplitude (or y or vertical) variability. Tradi- tional methods often ignore the phase variability and focus solely on the amplitude variation, using cross-sectional techniques such as fPCA for dimensional reduction and data modeling. Ignoring phase variability leads to a loss of structure in the data and inefficiency in data models. This paper presents an approach that relies on separating the phase (x-axis) and amplitude (y-axis), then modeling these components using joint distributions. This separation, in turn, is performed using a technique called elastic shape analysis of curves that involves a new mathematical representation of functional data. Then, using individual fPCAs, one each for phase and amplitude components, while respecting the nonlinear geometry of the phase representation space; impose joint probability models on principal coefficients of these components. These ideas are demonstrated using random sampling, for models estimated from simulated and real datasets, and show their superiority over models that ignore phase-amplitude separation. Furthermore, the generative models are applied to classification of functional data and achieve high performance in applications involv- ing SONAR signals of underwater objects, handwritten signatures, and periodic body movements recorded by smart phones.

stat.CO↗