arXiv subjects

Amanda Lenzi

Publications and source records attributed to Amanda Lenzi.

11 recordsLinked to original sources

arXiv2025

A sub-hourly spatio-temporal statistical model for solar irradiance in Ireland using open-source data

Accurate estimation of solar irradiance is essential for reliable modelling of solar photovoltaic (PV) power production. In Ireland's highly variable maritime climate, where ground-based measurement stations are sparsely distributed, selecting an appropriate solar irradiance dataset presents a significant challenge. This study introduces a novel Bayesian spatio-temporal modelling framework for predicting solar irradiance at hourly and sub-hourly (10-minute) resolutions across Ireland. Cross-validation demonstrates that our model is statistically robust across all temporal resolutions with hourly showing highest prediction precision whereas 10-minute resolution encounters higher errors but better uncertainty quantification. In separate evaluations, we compare our model against alternative data sources, including reanalysis datasets and nearest-station interpolation, and find that it consistently provides superior site-specific accuracy. At the hourly scale, our model outperforms ERA5 in agreement with ground-based observations. At the sub-hourly scale, 10-minute resolution estimates provide solar PV power outputs consistent with residential and industrial solar PV installations in Ireland. Beyond surpassing existing datasets, our model delivers full uncertainty quantification, scalability and the capacity for real-time implementation, offering a powerful tool for solar energy prediction and the estimation of losses due to overload clipping from inverter undersizing.

stat.AP↗

arXiv2024

When the whole is greater than the sum of its parts: Scaling black-box inference to large data settings through divide-and-conquer

Black-box methods such as deep neural networks are exceptionally fast at obtaining point estimates of model parameters due to their amortisation of the loss function computation, but are currently restricted to settings for which simulating training data is inexpensive. When simulating data is computationally expensive, both the training and uncertainty quantification, which typically relies on a parametric bootstrap, become intractable. We propose a black-box divide-and-conquer estimation and inference framework when data simulation is computationally expensive that trains a black-box estimation method on a partition of the multivariate data domain, estimates and bootstraps on the partitioned data, and combines estimates and inferences across data partitions. Through the divide step, only small training data need be simulated, substantially accelerating the training. Further, the estimation and bootstrapping can be conducted in parallel across multiple computing nodes to further speed up the procedure. Finally, the conquer step accounts for any dependence between data partitions through a statistically and computationally efficient weighted average. We illustrate the implementation of our framework in high-dimensional spatial settings with Gaussian and max-stable processes. Applications to modeling extremal temperature data from both a climate model and observations from the National Oceanic and Atmospheric Administration highlight the feasibility of estimation and inference of max-stable process parameters with tens of thousands of locations.

stat.ME↗

arXiv2024

Demonstrating the power and flexibility of variational assumptions for amortized neural posterior estimation in environmental applications

Classic Bayesian methods with complex models are frequently infeasible due to an intractable likelihood. Simulation-based inference methods, such as Approximate Bayesian Computing (ABC), calculate posteriors without accessing a likelihood function by leveraging the fact that data can be quickly simulated from the model, but converge slowly and/or poorly in high-dimensional settings. In this paper, we propose a framework for Bayesian posterior estimation by mapping data to posteriors of parameters using a neural network trained on data simulated from the complex model. Posterior distributions of model parameters are efficiently obtained by feeding observed data into the trained neural network. We show theoretically that our posteriors converge to the true posteriors in Kullback-Leibler divergence. Our approach yields computationally efficient and theoretically justified uncertainty quantification, which is lacking in existing simulation-based neural network approaches. Comprehensive simulation studies highlight our method's robustness and accuracy.

stat.CO↗

arXiv2023

Neural Likelihood Surfaces for Spatial Processes with Computationally Intensive or Intractable Likelihoods

In spatial statistics, fast and accurate parameter estimation, coupled with a reliable means of uncertainty quantification, can be challenging when fitting a spatial process to real-world data because the likelihood function might be slow to evaluate or wholly intractable. In this work, we propose using convolutional neural networks to learn the likelihood function of a spatial process. Through a specifically designed classification task, our neural network implicitly learns the likelihood function, even in situations where the exact likelihood is not explicitly available. Once trained on the classification task, our neural network is calibrated using Platt scaling which improves the accuracy of the neural likelihood surfaces. To demonstrate our approach, we compare neural likelihood surfaces and the resulting maximum likelihood estimates and approximate confidence regions with the equivalent for exact or approximate likelihood for two different spatial processes: a Gaussian process and a Brown-Resnick process which have computationally intensive and intractable likelihoods, respectively. We conclude that our method provides fast and accurate parameter estimation with a reliable method of uncertainty quantification in situations where standard methods are either undesirably slow or inaccurate. The method is applicable to any spatial process on a grid from which fast simulations are available.

stat.ME↗

arXiv2023

Towards black-box parameter estimation

Deep learning algorithms have recently shown to be a successful tool in estimating parameters of statistical models for which simulation is easy, but likelihood computation is challenging. But the success of these approaches depends on simulating parameters that sufficiently reproduce the observed data, and, at present, there is a lack of efficient methods to produce these simulations. We develop new black-box procedures to estimate parameters of statistical models based only on weak parameter structure assumptions. For well-structured likelihoods with frequent occurrences, such as in time series, this is achieved by pre-training a deep neural network on an extensive simulated database that covers a wide range of data sizes. For other types of complex dependencies, an iterative algorithm guides simulations to the correct parameter region in multiple rounds. These approaches can successfully estimate and quantify the uncertainty of parameters from non-Gaussian models with complex spatial and temporal dependencies. The success of our methods is a first step towards a fully flexible automatic black-box estimation framework.

stat.ML↗

arXiv2022

Spatio-Temporal Cross-Covariance Functions under the Lagrangian Framework with Multiple Advections

When analyzing the spatio-temporal dependence in most environmental and earth sciences variables such as pollutant concentrations at different levels of the atmosphere, a special property is observed: the covariances and cross-covariances are stronger in certain directions. This property is attributed to the presence of natural forces, such as wind, which cause the transport and dispersion of these variables. This spatio-temporal dynamics prompted the use of the Lagrangian reference frame alongside any Gaussian spatio-temporal geostatistical model. Under this modeling framework, a whole new class was birthed and was known as the class of spatio-temporal covariance functions under the Lagrangian framework, with several developments already established in the univariate setting, in both stationary and nonstationary formulations, but less so in the multivariate case. Despite the many advances in this modeling approach, efforts have yet to be directed to probing the case for the use of multiple advections, especially when several variables are involved. Accounting for multiple advections would make the Lagrangian framework a more viable approach in modeling realistic multivariate transport scenarios. In this work, we establish a class of Lagrangian spatio-temporal cross-covariance functions with multiple advections, study its properties, and demonstrate its use on a bivariate pollutant dataset of particulate matter in Saudi Arabia.

stat.ME↗

arXiv2021

Neural Networks for Parameter Estimation in Intractable Models

We propose to use deep learning to estimate parameters in statistical models when standard likelihood estimation methods are computationally infeasible. We show how to estimate parameters from max-stable processes, where inference is exceptionally challenging even with small datasets but simulation is straightforward. We use data from model simulations as input and train deep neural networks to learn statistical parameters. Our neural-network-based method provides a competitive alternative to current approaches, as demonstrated by considerable accuracy and computational time improvements. It serves as a proof of concept for deep learning in statistical parameter estimation and can be extended to other estimation problems.

stat.ME↗

arXiv2021

Parameter Estimation with Dense and Convolutional Neural Networks Applied to the FitzHugh-Nagumo ODE

Machine learning algorithms have been successfully used to approximate nonlinear maps under weak assumptions on the structure and properties of the maps. We present deep neural networks using dense and convolutional layers to solve an inverse problem, where we seek to estimate parameters of a FitzHugh-Nagumo model, which consists of a nonlinear system of ordinary differential equations (ODEs). We employ the neural networks to approximate reconstruction maps for model parameter estimation from observational data, where the data comes from the solution of the ODE and takes the form of a time series representing dynamically spiking membrane potential of a biological neuron. We target this dynamical model because of the computational challenges it poses in an inference setting, namely, having a highly nonlinear and nonconvex data misfit term and permitting only weakly informative priors on parameters. These challenges cause traditional optimization to fail and alternative algorithms to exhibit large computational costs. We quantify the prediction errors of model parameters obtained from the neural networks and investigate the effects of network architectures with and without the presence of noise in observational data. We generalize our framework for neural network-based reconstruction maps to simultaneously estimate ODE parameters and parameters of autocorrelated observational noise. Our results demonstrate that deep neural networks have the potential to estimate parameters in dynamical models and stochastic processes, and they are capable of predicting parameters accurately for the FitzHugh-Nagumo model.

stat.ML↗

arXiv2020

Improving Bayesian Local Spatial Models in Large Data Sets

Environmental processes resolved at a sufficiently small scale in space and time will inevitably display non-stationary behavior. Such processes are both challenging to model and computationally expensive when the data size is large. Instead of modeling the global non-stationarity explicitly, local models can be applied to disjoint regions of the domain. The choice of the size of these regions is dictated by a bias-variance trade-off; large regions will have smaller variance and larger bias, whereas small regions will have higher variance and smaller bias. From both the modeling and computational point of view, small regions are preferable to better accommodate the non-stationarity. However, in practice, large regions are necessary to control the variance. We propose a novel Bayesian three-step approach that allows for smaller regions without compromising the increase of the variance that would follow. We are able to propagate the uncertainty from one step to the next without issues caused by reusing the data. The improvement in inference also results in improved prediction, as our simulated example shows. We illustrate this new approach on a data set of simulated high-resolution wind speed data over Saudi Arabia.

stat.AP↗

arXiv2017

Benefits of spatio-temporal modelling for short term wind power forecasting at both individual and aggregated levels

The share of wind energy in total installed power capacity has grown rapidly in recent years around the world. Producing accurate and reliable forecasts of wind power production, together with a quantification of the uncertainty, is essential to optimally integrate wind energy into power systems. We build spatio-temporal models for wind power generation and obtain full probabilistic forecasts from 15 minutes to 5 hours ahead. Detailed analysis of the forecast performances on the individual wind farms and aggregated wind power are provided. We show that it is possible to improve the results of forecasting aggregated wind power by utilizing spatio-temporal correlations among individual wind farms. Furthermore, spatio-temporal models have the advantage of being able to produce spatially out-of-sample forecasts. We evaluate the predictions on a data set from wind farms in western Denmark and compare the spatio-temporal model with an autoregressive model containing a common autoregressive parameter for all wind farms, identifying the specific cases when it is important to have a spatio-temporal model instead of a temporal one. This case study demonstrates that it is possible to obtain fast and accurate forecasts of wind power generation at wind farms where data is available, but also at a larger portfolio including wind farms at new locations. The results and the methodologies are relevant for wind power forecasts across the globe as well as for spatial-temporal modelling in general.

stat.AP↗

arXiv2014

Analysis of Aggregated Functional Data from Mixed Populations with Application to Energy Consumption

Understanding the energy consumption patterns of different types of consumers is essential in any planning of energy distribution. However, obtaining consumption information for single individuals is often either not possible or too expensive. Therefore, we consider data from aggregations of energy use, that is, from sums of individuals' energy use, where each individual falls into one of C consumer classes. Unfortunately, the exact number of individuals of each class may be unknown: consumers do not always report the appropriate class, due to various factors including differential energy rates for different consumer classes. We develop a methodology to estimate the expected energy use of each class as a function of time and the true number of consumers in each class. We also provide some measure of uncertainty of the resulting estimates. To accomplish this, we assume that the expected consumption is a function of time that can be well approximated by a linear combination of B-splines. Individual consumer perturbations from this baseline are modeled as B-splines with random coefficients. We treat the reported numbers of consumers in each category as random variables with distribution depending on the true number of consumers in each class and on the probabilities of a consumer in one class reporting as another class. We obtain maximum likelihood estimates of all parameters via a maximization algorithm. We introduce a special numerical trick for calculating the maximum likelihood estimates of the true number of consumers in each class. We apply our method to a data set and study our method via simulation.

stat.AP↗