Worldwide epidemics, such as corona virus disease 2019 (COVID-19), cause unprecedented challenges for society and its healthcare systems. Governments attempt to mitigate those challenges by either reducing healthcare demand (“flattening the curve” by imposing restrictions, e.g., on travel or social gatherings) or by increasing healthcare capacity, for example, by canceling elective procedures or setting up field hospitals. To implement these mitigation procedures efficiently, accurate and timely forecasts of the epidemic’s progression are necessary. In this paper, we develop an innovative forecasting methodology based on the ideas of long short-term memory (LSTM) recurrent neural networks. LSTM models are shown to outperform traditional forecasting models, especially when the relationship between input and output is complex and not available in closed form. However, whereas LSTM models perform well for data that changes dynamically over time, one shortcoming is that they are not directly applicable when the data also includes static, nontemporal components. In this work, we propose an [Formula: see text] model that overcomes this limitation. Our model leverages a private partnership with a mobile data company in order to capture population mobility (using mobility indices derived from mobile device data), which allows us to anticipate an epidemic’s spread early and accurately. In addition, we also leverage a public partnership with a consortium of hospitals. Using hospital admissions (rather than, say, positive caseload) results in an unbiased measure of the severity of an epidemic because patients seek and are admitted to hospital care only when symptoms worsen beyond a critical point. We illustrate the effectiveness of our method on forecasting COVID-19 for a major U.S. metropolitan area where it has aided decision makers of the emergency policy group. Our model improves the predictive accuracy of hospital admission by a factor of 2.5× as compared with competing models in the same analytical space. History: Accepted by J. Paul Brooks, Area Editor for Applications in Biology, Medicine, & Healthcare. Funding: This research was funded by a monetary gift from Hillsborough County to establish the Pandemic Response Research Fund at University of South Florida. Supplemental Material: The software that supports the findings of this study is available within the paper and its Supplemental Information ( https://pubsonline.informs.org/doi/suppl/10.1287/ijoc.2023.1269 ) as well as from the IJOC GitHub software repository ( https://github.com/INFORMSJoC/2021.0027 ) at ( http://dx.doi.org/10.5281/zenodo.7112004 ).

The flexibility of time and location as well as the availability of an abundance of both old and new products makes online auctions an important part of people's daily shopping experience. Whereas many bidders rely on variants of the well-documented early or last-minute bidding strategies, neither strategy takes into account the aspect of auction competition: at any point in time, there are hundreds, even thousands, of the same or similar items up for sale, competing for the same bidder. In this paper, we propose a novel automated and data-driven bidding strategy. Our strategy consists of two main components. First, we develop a dynamic, forward-looking model for price in competing auctions. By incorporating dynamic features of the auction process and its competitive environment, our model is capable of accurately predicting an auction's price and outperforming model alternatives such as the generalized additive model, classification and regression trees, or Neural Networks. Then, using the idea of maximizing a bidder's surplus, we build a bidding framework around this model that selects the best auction to bid on and determines the best bid amount. The best auction is given by the one that yields the highest predicted surplus; the best bid amount is given by its predicted auction price. Our approach maximizes expected surplus and balances the probability of winning an auction with its average surplus. In simulations, we compare our automated strategy with early and last-minute bidding and find that our approach extracts 97% and 15% more expected surplus, respectively.

Bids during an online auction arrive at unequally-spaced discrete time points. Our goal is to capture the entire continuous price-evolution function by representing it as a functional object. Various nonparametric smoothing methods exist to recover the functional object from the observed discrete bid data. Previous studies use penalized polynomial and monotone smoothing splines; however, these require the determination and storage of a large number of coefficients and often lengthy computational time. We present a family of parametric growth curves that describe the price-evolution during online auctions. This approach is parsimonious and has an appealing interpretation in the online auction context. We also provide an automated fitting algorithm that is computationally fast. Methods are illustrated using eBay data.

Time-series forecasting has a large number of applications. Users with a partial time series for auctions, new stock offerings, or industrial processes desire estimates of the future behavior. We present a data driven forecasting method and interface called similarity-based forecasting (SBF). A pattern matching search in an historical time series dataset produces a subset of curves similar to the partial time series. The forecast is displayed graphically as a river plot showing statistical information about the SBF subset. A forecasting preview interface allows users to interactively explore alternative pattern matching parameters and see multiple forecasts simultaneously. User testing with 8 users demonstrated advantages and led to improvements.

We propose a dynamic model for forecasting price in online auctions. One of the key features of our model is that it operates during the live-auction, which makes it different from previous approaches that only consider static models. Our model is also different with respect to how information about price is incorporated. While one part of the model is based on the more traditional notion of an auction's price-level, another part incorporates its dynamics in the form of a price's velocity and acceleration. In that sense, it incorporates key features of a dynamic environment such as an online auction. The use of novel functional data methodology allows us to measure, and subsequently include, dynamic price characteristics. We illustrate our model on a diverse set of eBay auctions across many different book categories. We find significantly higher prediction accuracy compared to standard approaches.