Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Statistical forecasting”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 37 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Use of Physics to Improve Solar Forecast: Physics-Informed Persistence Models for Simultaneously Forecasting GHI, DNI, and DHI

Observation-based statistical models have been widely used in forecasting solar energy; however, existing models often lack a clear relation to physics and are limited largely to global horizontal irradiance (GHI) forecasts over relatively short time horizons (< 1 hour). Incorporating physics into observation-based models, increasing forecast time horizons and developing a model system for forecasting not only GHI but also direct normal irradiance (DNI) and diffuse horizontal irradiance (DHI) remain challenging, especially under cloudy conditions because of complex cloud-radiation interactions. This work attempts to address these challenges by developing a hierarchy of four new physics-informed persistence models that can be used to simultaneously forecast GHI, DNI and DHI. The decade-long measurements (1998 to 2014) at the U.S. Department of Energy's Atmospheric Radiation Measurement (ARM)'s Southern Great Plains (SGP) Central Facility site are used to evaluate the performance of the new models. Overall, the results show that the new physics-informed forecast models generally outperform the simple and smart persistence models, and improve the forecast accuracy at lead times from 1.25 hours up to 6 hours. Further analysis reveals that the forecast error is highly related to the error and temporal variability of the assumed cloud predictor. The best model for forecasting different radiative components can be explained by the relationship between solar irradiances and cloud properties.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Improved Subseasonal Forecasting of Extreme Polar Vortices Using Machine Learning

Our research was focused on forecasting the position and shape of the winter stratospheric polar vortex at a subseasonal timescale of 15 days in advance. To achieve this, we employed both statistical and neural network machine learning techniques. The analysis was performed on 42 winter seasons of reanalysis data provided by NASA giving us a total of 6,342 days of data. The state of the polar vortex for determined by using geometric moments to calculate the centroid latitude and the aspect ratio of an ellipse fit onto the vortex. Timeseries for thirty additional precursors were calculated to help improve the predictive capabilities of the algorithm. Feature importance of these precursors was performed using random forest to measure the predictive importance and the ideal number of precursors. Then, using the precursors identified as important, various statistical methods were tested for predictive accuracy with random forest and nearest neighbor performing the best. An echo state network, a type of recurrent neural network that features sparsely connected hidden layer and a reduced number of trainable parameters that allows for rapid training and testing, was also implemented for the forecasting problem. Hyperparameter tuning was performed for each methods using a subset of the training data. The algorithms were trained and tuned on the first 41 years of data, then tested for accuracy on the final year. In general, the centroid latitude of the polar vortex proved easier to predict than the aspect ratio across all algorithms. Random forest outperformed other statistical forecasting algorithms overall but struggled to predict extreme values. Forecasting from echo state network suggested a strong predictive capability past 15 days, but further work is required to fully realize the potential of recurrent neural network approaches.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Data-Driven Transition Path Analysis Yields a Statistical Understanding of Sudden Stratospheric Warming Events in an Idealized Model

Abstract Atmospheric regime transitions are highly impactful as drivers of extreme weather events, but pose two formidable modeling challenges: predicting the next event (weather forecasting) and characterizing the statistics of events of a given severity (the risk climatology). Each event has a different duration and spatial structure, making it hard to define an objective “average event.” We argue here that transition path theory (TPT), a stochastic process framework, is an appropriate tool for the task. We demonstrate TPT’s capacities on a wave–mean flow model of sudden stratospheric warmings (SSWs) developed by Holton and Mass, which is idealized enough for transparent TPT analysis but complex enough to demonstrate computational scalability. Whereas a recent article (Finkel et al. 2021) studied near-term SSW predictability, the present article uses TPT to link predictability to long-term SSW frequency. This requires not only forecasting forward in time from an initial condition, but also backward in time to assess the probability of the initial conditions themselves. TPT enables one to condition the dynamics on the regime transition occurring, and thus visualize its physical drivers with a vector field called the reactive current . The reactive current shows that before an SSW, dissipation and stochastic forcing drive a slow decay of vortex strength at lower altitudes. The response of upper-level winds is late and sudden, occurring only after the transition is almost complete from a probabilistic point of view. This case study demonstrates that TPT quantities, visualized in a space of physically meaningful variables, can help one understand the dynamics of regime transitions.

Meteorology & Atmospheric Sciences↗
DOE OSTI2022

A High‐Resolution Tropical Mesoscale Convective System Reanalysis (TMeCSR)

Abstract Modern global reanalysis products have greatly accelerated meteorological research in synoptic‐to‐planetary‐scale phenomena. However, their use in studying tropical mesoscale convective systems (MCSs) and their regional‐to‐global impact has mostly been limited to supplying initial and boundary conditions for MCS‐resolving simulations and providing information about the large‐scale environments of MCSs. These limitations are due to difficulties in resolving tropical MCS dynamics in the relatively low‐resolution global models and that tropical MCSs often occur over poorly observed regions. In this work, a Tropical MCS‐resolving Reanalysis product (TMeCSR) was created over a region with frequent tropical MCSs. This region spans the tropical Indian Ocean, tropical continental Asia, Maritime Continent, and Western Pacific. TMeCSR is produced by assimilating all‐sky infrared radiances from geostationary satellites and other conventional observations into an MCS‐resolving regional model using the Ensemble Kalman Filter. The resulting observation‐constrained high‐resolution (9‐km grid spacing) data set is available hourly during the boreal summer (June‐August) of 2017, during which widespread severe flooding occurred. Comparisons of TMeCSR and European Center for Medium Range Weather Forecast Reanalysis version 5 (ERA5) against independent satellite retrievals indicate that TMeCSR's cloud and multiscale rain fields are better than those of ERA5. Furthermore, TMeCSR better captured the diurnal variability of rainfall and the statistical characteristics of MCSs. Forecasts initialized from TMeCSR also have more accurate rain and clouds than those initialized from ERA5. The TMeCSR and ERA5 forecasts have similar performances with respect to sounding and surface observations. These results indicate that TMeCSR is a promising MCS‐resolving data set for tropical MCS studies.

Chan, Man‐Yau↗
DOE OSTI2025

Improving tropical cyclone rapid intensification forecasts with satellite measurements of sea surface salinity and calibrated machine learning

Forecasting rapid intensification (RI) of tropical cyclones (TC) is a mission known for large errors. One under-researched factor that affects TC intensification is salinity, which is important for density stratification in certain ocean regions and can affect the surface enthalpy flux under a strengthening hurricane. To investigate the impact and efficacy of using salinity information in state-of-the-art forecasting, we use a statistical model consisting of a variety of machine learning (ML) methods. For salinity data, we use satellite measurements of pre-storm sea surface salinity (SSS) as a proxy for the salinity stratification. We train and test the model on various ocean basins, including the Atlantic, eastern North Pacific and western North Pacific. A calibrator is trained on top of the ML models to correct and enhance probability forecasts. The calibrator significantly improves probability forecasts relative to recent works. The ML model performance is improved with the addition of SSS in the Eastern North Pacific, western North Pacific, and the Caribbean subregion of the North Atlantic, and the overall model performance is better than previous studies. SSS decreases model skill for a model trained on the full Atlantic basin. In the Indian Ocean, SSS is also notably correlated with RI occurrence, but the TC samples are not sufficient to train ML models.

hurricane↗
DOE OSTI2026

Short-term electricity load forecasting: Application-driven evaluation of machine learning models across spatial and temporal scales

As we transition towards a decarbonized economy, the integration of variable renewable energy resources and new demands (e.g., electric vehicles, heat pumps) into the electricity grid places unprecedented pressure on grid operators to effectively anticipate and manage peak load. In this context, machine learning algorithms are proving to be indispensable for accurate short-term load forecasting, a crucial task to address these challenges. This study benchmarks 6 machine learning algorithms, including three neural networks and three tree-based algorithms, across various levels of spatial aggregation and time horizons (1, 4, 8, 24, and 48 h). The central contribution of this work is the comparison and analysis of load forecasting models not only based on statistical metrics, but also based on a novel error metric, which evaluates the cost implications of forecast errors for power system stakeholders. Results show that tree-based models outperform neural networks, based on statistical metrics, and yield less skewed error distributions for most spatial scales. However, through the lens of the novel error metric, neural networks are the more competitive choice, especially for forecast horizons that exceed 8 h. The study concludes with actionable recommendations to grid operators and highlights the need for the development of error metrics that link forecasting accuracy to operational costs. To promote transparency and open science, the datasets and Python code are open-sourced via a supplementary repository.

Houben, Nikolaus↗
DOE OSTI2022

Incorporate day-ahead robustness and real-time incentives for electricity market design

In this paper, we propose a two-stage electricity market framework to explore the participation of distributed energy resources (DERs) in a day-ahead (DA) market and a real-time (RT) market. The objective is to determine the optimal bidding strategies of the aggregated DERs in the DA market and generate online incentive signals for DER-owners to optimize the social-welfare taking into account network operational constraints. Distributionally robust optimization is used to explicitly incorporate data-based statistical information of renewable forecasts into the supply/demand decisions in the DA market. We evaluate the conservativeness of bidding strategies distinguished by different risk aversion settings. In the RT market, a bi-level time-varying optimization problem is proposed to design the online incentive signals to tradeoff the RT imbalance penalty for distribution system operators (DSOs) and the costs of individual DER-owners. This enables tracking their optimal dispatch to provide fast balancing services, in the presence of time-varying network states while satisfying the voltage regulation requirement. Simulation results on both DA wholesale market and RT balancing market demonstrate the necessity of this two-stage design, and its robustness to uncertainties, the performance of convergence, the tracking ability and the feasibility of the resulting network operations.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2023

Short-lead seasonal precipitation forecast in northeastern Brazil using an ensemble of artificial neural networks

This study assesses the deterministic and probabilistic forecasting skill of a 1-month-lead ensemble of Artificial Neural Networks (EANN) based on low-frequency climate oscillation indices. The predictand is the February-April (FMA) rainfall in the Brazilian state of Ceará, which is a prominent subject in climate forecasting studies due to its high seasonal predictability. Additionally, the study proposes combining the EANN with dynamical models into a hybrid multi-model ensemble (MME). The forecast verification is carried out through a leave-one-out cross-validation based on 40 years of data. The EANN forecasting skill is compared with traditional statistical models and the dynamical models that compose Ceará’s operational seasonal forecasting system. A spatial comparison showed that the EANN was among the models with the smallest Root Mean Squared Error (RMSE) and Ranked Probability Score (RPS) in most regions. Moreover, the analysis of the area-aggregated reliability showed that the EANN is better calibrated than the individual dynamical models and has better resolution than Multinomial Logistic Regression for above-normal (AN) and below-normal (BN) categories. It is also shown that combining the EANN and dynamical models into a hybrid MME reduces the overconfidence of the extreme categories observed in a dynamically-based MME, improving the reliability of the forecasting system.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Descriptors of water aggregation

For this work, we rely on a total of 23 (cluster size, 8 structural, and 14 connectivity) descriptors to investigate structural patterns and connectivity motifs associated with water cluster aggregation. In addition to the cluster size n (number of molecules), the 8 structural descriptors can be further categorized into (i) one-body (intramolecular): covalent OH bond length (r OH ) and HOH bond angle (θ HOH ), (ii) two-body: OO distance (r OO ), OHO angle (θ OHO ), and HOOX dihedral angle ($\phi$ HOOX ), where X lies on the bisector of the HOH angle, (iii) three-body: OOO angle (θ OOO ), and (iv) many-body: modified tetrahedral order parameter (q) to account for two-, three-, four-, five-coordinated molecules (q m , m = 2, 3, 4, 5) and radius of gyration (R g ). The 14 connectivity descriptors are all many-body in nature and consist of the AD, AAD, ADD, AADD, AAAD, AAADD adjacencies [number of hydrogen bonds accepted (A) and donated (D) by each water molecule], Wiener index, Average Shortest Path Length, hydrogen bond saturation (% HB), and number of non-short-circuited three-membered cycles, four-membered cycles, five-membered cycles, six-membered cycles, and seven-membered cycles. We mined a previously reported database of 4 948 959 water cluster minima for (H 2 O) n , n = 3–25 to analyze the evolution and correlation of these descriptors for the clusters within 5 kcal/mol of the putative minima. It was found that r OH and % HB correlated strongly with cluster size n, which was identified as the strongest predictor of energetic stability. Marked changes in the adjacencies and cycle count were observed, lending insight into changes in the hydrogen bond network upon aggregation. A Principal Component Analysis (PCA) was employed to identify descriptor dependencies and group clusters into specific structural patterns across different cluster sizes. The results of this study inform our understanding of how water clusters evolve in size and what appropriate descriptors of their structural and connectivity patterns are with respect to system size, stability, and similarity. The approach described in this study is general and can be easily extended to other hydrogen-bonded systems.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

Error-Level-Controlled Synthetic Forecasts for Renewable Generation

Renewable energy resources, including solar and wind energy, play a significant role in sustainable energy systems. However, the inherent uncertainty and intermittency of renewable generation pose challenges to the safe and efficient operation of power systems. Recognizing the importance of short-term (hours ahead) renewable generation forecasting in power systems operation, it becomes crucial to address the potential inaccuracies in these forecasts. To systematically evaluate the performance of controllers in the presence of imperfect forecasts, we generate synthetic forecasts using actual renewable generation profiles (one from solar and one from wind). These synthetic forecasts incorporate different levels of statistical error, allowing us to control and manipulate the accuracy of the predictions. The primary objective is to employ synthetic forecasts with controlled yet realistic error levels to systematically investigate how controllers adapt to variations in forecast accuracy, providing valuable insights into their robustness and effectiveness under real-world conditions.

Array↗
DOE OSTI2021

Deep Learning Experiments for Tropical Cyclone Intensity Forecasts

Reducing tropical cyclone (TC) intensity forecast errors is a challenging task that has interested the operational forecasting and research community for decades. To address this, we developed a deep learning (DL)-based multilayer perceptron (MLP) TC intensity prediction model. The model was trained using the global Statistical Hurricane Intensity Prediction Scheme (SHIPS) predictors to forecast the change in TC maximum wind speed for the Atlantic basin. In the first experiment, a 24-h forecast period was considered. To overcome sample size limitations, we adopted a leave one year out (LOYO) testing scheme, where a model is trained using data from all years except one and then evaluated on the year that is left out. When tested on 2010–18 operational data using the LOYO scheme, the MLP outperformed other statistical–dynamical models by 9%–20%. Additional independent tests in 2019 and 2020 were conducted to simulate real-time operational forecasts, where the MLP model again outperformed the statistical–dynamical models by 5%–22% and achieved comparable results as HWFI. The MLP model also correctly predicted more rapid intensification events than all the four operational TC intensity models compared. In the second experiment, we developed a lightweight MLP for 6-h intensity predictions. When coupled with a synthetic TC track model, the lightweight MLP generated realistic TC intensity distribution in the Atlantic basin. Therefore, the MLP-based approach has the potential to improve operational TC intensity forecasts, and will also be a viable option for generating synthetic TCs for climate studies.

58 GEOSCIENCES↗
DOE OSTI2020

Understanding Differences in California Climate Projections Produced by Dynamical and Statistical Downscaling

Abstract We compare historical and end‐of‐century temperature and precipitation patterns over California from one dynamically downscaled simulation using the Weather Research and Forecast (WRF) model and two simulations statistically downscaled using Localized Constructed Analogs (LOCA). We uniquely separate causes of differences between dynamically and statistically based future climate projections into differences in historical climate (gridded observations versus regional climate model output) and differences in how these downscaling techniques explicitly handle future climate changes (numerical modeling versus analogs). In these methods, solutions between different downscaling techniques differ more in the future compared to the historical period. Changes projected by LOCA are insensitive to the choice of driving data. Only through dynamical downscaling can we simulate physically consistent regional springtime warming patterns across the Sierra Nevada, while the statistical simulations inherit an unphysical signal from their parent Global Climate Model (GCM) or gridded data. The results of our study clarify why these different techniques produce different outcomes and may also provide guidance on which downscaled products to use for certain impact analyses in California and perhaps other Mediterranean regimes.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Redshift evolution and covariances for joint lensing and clustering studies with DESI Y1

ABSTRACT Galaxy–galaxy lensing (GGL) and clustering measurements from the Dark Energy Spectroscopic Instrument Year 1 (DESI Y1) data set promise to yield unprecedented combined-probe tests of cosmology and the galaxy–halo connection. In such analyses, it is essential to identify and characterize all relevant statistical and systematic errors. We forecast the covariances of DESI Y1 GGL + clustering measurements and the systematic bias due to redshift evolution in the lens samples. Focusing on the projected clustering and GGL correlations, we compute a Gaussian analytical covariance, using a suite of N-body and lognormal simulations to characterize the effect of the survey footprint. Using the DESI one percent survey data, we measure the evolution of galaxy bias parameters for the DESI luminous red galaxy (LRG) and bright galaxy survey (BGS) samples. We find mild evolution in the LRGs in $0.4 < z < 0.8$, subdominant to the expected statistical errors. For BGS, we find less evolution for brighter absolute magnitude cuts, at the cost of reduced sample size. We find that for a redshift bin width $\Delta z = 0.1$, evolution effects on DESI Y1 GGL is negligible across all scales, all fiducial selection cuts, all fiducial redshift bins. Galaxy clustering is more sensitive to evolution due to the bias squared scaling. Nevertheless the redshift evolution effect is insignificant for clustering above the 1-halo scale of $0.1h^{-1}$ Mpc. For studies that wish to reliably access smaller scales, additional treatment of redshift evolution is likely needed. This study serves as a reference for GGL and clustering studies using the DESI Y1 sample.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2023

A hybrid data-driven and model-based approach for computationally efficient stochastic unit commitment and economic dispatch under wind and solar uncertainty

Stochastic unit commitment (UC) and economic dispatch (ED) are imperative in dealing with uncertainty in renewable forecast for power system operation and planning such that the overall expected production cost is minimized over the planning horizon. However, accurate calculation of the expected production cost requires assessment of a very large number of different scenarios of uncertain renewable resources, such as solar and wind, which is practically infeasible to simulate in real time. This article proposes a hybrid datadriven and physics-based model-predictive paradigm to efficiently solve for stochastic unit commitment and economic dispatch considering uncertainty in wind and solar power forecasts. Here, the novelty of the approach lies in decoupling the production cost estimation from the unit commitment and economic dispatch optimization problems under uncertainty without compromising on the fidelity of the solutions. A data-driven machine learning model is first developed to predict the mean optimal production cost. A physics-based inverse problem is then solved to get the stochastic UC and ED profiles from the expected cost. The presented approach considers, for the first time, solar uncertainty in UC/ED determination and enables efficient and accurate propagation of wind and solar uncertainty to estimate the statistics of the production cost. The effectiveness of the developed approach is demonstrated systematically on a stylized RTS-GMLC single-node system. The overall framework predicts the expected cost 62.5% more accurately than the existing state-of-the-art, on unforeseen days during the entire year, and yields, for the first time, the associated physically consistent UC and ED profiles. The solutions are also shown to be flexible in providing adequate daily reserves to address any statistical deviations from probabilistic power forecasts. The computational time associated with the presented method is only about 10 s compared to over 24 h needed for a conventional stochastic UC/ED determination under uncertainty on an Intel Core i9 processor with 32 GB of RAM.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2021

Review of Onsite Temperature and Solar Forecasting Models to Enable Better Building Design and Operations

Advanced building controls and energy optimization for new constructions and retrofits rely on accurate weather data. Traditionally, most studies utilize airport weather information as the decision inputs. However, most buildings are in environments that are quite different than those at the airport miles away. Tree cover, adjacent buildings, and micro-climate effects caused by the larger surrounding area can all yield deviations in air temperature, humidity, solar irradiance, and wind that are large enough to influence design and operation decisions. In order to overcome this challenge, there are many prior studies on developing weather forecasting algorithms from micro-to meso-scales. Additionally, this paper reviews and complies knowledge on common weather data resources, data processing methodologies and forecasting techniques of weather information. Commonly used statistical, machine learning and physical-based models are discussed and presented as two major categories: deterministic forecasting and probabilistic forecasting. Finally, evaluation metrics for forecasting errors are listed and discussed.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2024

Systematic Effects in Galaxy-Galaxy Lensing with DESI

The Dark Energy Spectroscopic Instrument (DESI) survey will measure spectroscopic redshifts for millions of galaxies across roughly $14,000 \, \mathrm{deg}^2$ of the sky. Cross-correlating targets in the DESI survey with complementary imaging surveys allows us to measure and analyze shear distortions caused by gravitational lensing in unprecedented detail. In this work, we analyze a series of mock catalogs with ray-traced gravitational lensing and increasing sophistication to estimate systematic effects on galaxy-galaxy lensing estimators such as the tangential shear $\gamma_{\mathrm{t}}$ and the excess surface density $\Delta\Sigma$ . We employ mock catalogs tailored to the specific imaging surveys overlapping with the DESI survey: the Dark Energy Survey (DES), the Hyper Suprime-Cam (HSC) survey, and the Kilo-Degree Survey (KiDS). Among others, we find that fiber incompleteness can have significant effects on galaxy-galaxy lensing estimators but can be corrected effectively by up-weighting DESI targets with fibers by the inverse of the fiber assignment probability. Similarly, we show that intrinsic alignment and lens magnification are expected to be statistically significant given the precision forecasted for the DESI year-1 data set. Our study informs several analysis choices for upcoming cross-correlation studies of DESI with DES, HSC, and KiDS.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

Huge ensembles – Part 2: Properties of a huge ensemble of hindcasts generated with spherical Fourier neural operators

Abstract. In Part 1, we created an ensemble based on spherical Fourier neural operators. As initial condition perturbations, we used bred vectors, and as model perturbations, we used multiple checkpoints trained independently from scratch. Based on diagnostics that assess the ensemble's physical fidelity, our ensemble has comparable performance to operational weather forecasting systems. However, it requires orders-of-magnitude fewer computational resources. Here in Part 2, we generate a huge ensemble (HENS), with 7424 members initialized each day of summer 2023. We enumerate the technical requirements for running huge ensembles at this scale. HENS precisely samples the tails of the forecast distribution and presents a detailed sampling of internal variability. HENS has two primary applications: (1) as a large dataset with which to study the statistics and drivers of extreme weather and (2) as a weather forecasting system. For extreme climate statistics, HENS samples events 4σ away from the ensemble mean. At each grid cell, HENS increases the skill of the most accurate ensemble member and enhances coverage of possible future trajectories. As a weather forecasting model, HENS issues extreme weather forecasts with better uncertainty quantification. It also reduces the probability of outlier events, in which the verification value lies outside the ensemble forecast distribution.

Mahesh, Ankur↗
DOE OSTI2022

A Comparison of Infectious Disease Forecasting Methods across Locations, Diseases, and Time

Accurate infectious disease forecasting can inform efforts to prevent outbreaks and mitigate adverse impacts. This study compares the performance of statistical, machine learning (ML), and deep learning (DL) approaches in forecasting infectious disease incidences across different countries and time intervals. We forecasted three diverse diseases: campylobacteriosis, typhoid, and Q-fever, using a wide variety of features (n = 46) from public datasets, e.g., landscape, climate, and socioeconomic factors. We compared autoregressive statistical models to two tree-based ML models (extreme gradient boosted trees [XGB] and random forest [RF]) and two DL models (multi-layer perceptron and encoder–decoder model). The disease models were trained on data from seven different countries at the region-level between 2009–2017. Forecasting performance of all models was assessed using mean absolute error, root mean square error, and Poisson deviance across Australia, Israel, and the United States for the months of January through August of 2018. The overall model results were compared across diseases as well as various data splits, including country, regions with highest and lowest cases, and the forecasted months out (i.e., nowcasting, short-term, and long-term forecasting). Overall, the XGB models performed the best for all diseases and, in general, tree-based ML models performed the best when looking at data splits. There were a few instances where the statistical or DL models had minutely smaller error metrics for specific subsets of typhoid, which is a disease with very low case counts. Feature importance per disease was measured by using four tree-based ML models (i.e., XGB and RF with and without region name as a feature). The most important feature groups included previous case counts, region name, population counts and density, mortality causes of neonatal to under 5 years of age, sanitation factors, and elevation. This study demonstrates the power of ML approaches to incorporate a wide range of factors to forecast various diseases, regardless of location, more accurately than traditional statistical approaches.

59 BASIC BIOLOGICAL SCIENCES↗