Search for: All records

Abstract This article proposes a new statistical model to infer interpretable population-level preferences from ordinal comparison data. Such data is ubiquitous, e.g., ranked choice votes, top-10 movie lists, and pairwise sports outcomes. Traditional statistical inference on ordinal comparison data results in an overall ranking of objects, e.g., from best to worst, with each object having a unique rank. However, the ranks of some objects may not be statistically distinguishable. This could happen due to insufficient data or to the true underlying object qualities being equal. Because uncertainty communication in estimates of overall rankings is notoriously difficult, we take a different approach and allow groups of objects to have equal ranks or berank-clusteredin our model. Existing models related to rank-clustering are limited by their inability to handle a variety of ordinal data types, to quantify uncertainty, or by the need to pre-specify the number and size of potential rank-clusters. We solve these limitations through our proposed BayesianRank-Clustered Bradley–Terry–Luce (BTL)model. We accommodate rank-clustering via parameter fusion by imposing a novel spike-and-slab prior on object-specific worth parameters in the BTL family of distributions for ordinal comparisons. We demonstrate rank-clustering on simulated and real datasets in surveys, elections, and sports analytics. 

Abstract BackgroundIn many grant review settings, proposals are selected for funding on the basis of summary statistics of review ratings. Challenges of this approach (including the presence of ties and unclear ordering of funding preference for proposals) could be mitigated if rankings such as top-k preferences or paired comparisons, which are local evaluations that enforce ordering across proposals, were also collected and incorporated in the analysis of review ratings. However, analyzing ratings and rankings simultaneously has not been done until recently. This paper describes a practical method for integrating rankings and scores and demonstrates its usefulness for making funding decisions in real-world applications. MethodsWe first present the application of our existing joint model for rankings and ratings, the Mallows-Binomial, in obtaining an integrated score for each proposal and generating the induced preference ordering. We then apply this methodology to several theoretical “toy” examples of rating and ranking data, designed to demonstrate specific properties of the model. We then describe an innovative protocol for collecting rankings of the top-six proposals as an add-on to the typical peer review scoring procedures and provide a case study using actual peer review data to exemplify the output and how the model can appropriately resolve judges’ evaluations. ResultsFor the theoretical examples, we show how the model can provide a preference order to equally rated proposals by incorporating rankings, to proposals using ratings and only partial rankings (and how they differ from a ratings-only approach) and to proposals where judges provide internally inconsistent ratings/rankings and outlier scoring. Finally, we discuss how, using real world panel data, this method can provide information about funding priority with a level of accuracy in a well-suited format for research funding decisions. ConclusionsA methodology is provided to collect and employ both rating and ranking data in peer review assessments of proposal submission quality, highlighting several advantages over methods relying on ratings alone. This method leverages information to most accurately distill reviewer opinion into a useful output to make an informed funding decision and is general enough to be applied to settings such as in the NIH panel review process. 

Ecological and evolutionary dynamics are intrinsically entwined. On short timescales, ecological interactions determine the fate and impact of new mutants, while on longer timescales evolution shapes the entire community. Here, we study the evolution of large numbers of closely related strains with generalized Lotka Volterra interactions but no niche structure. Host-pathogen-like interactions drive the community into a spatiotemporally chaotic state characterized by continual, spatially-local, blooms and busts. Upon the slow serial introduction of new strains, the community diversifies indefinitely, accommodating an arbitrarily large number of strains in spite of the absence of stabilizing niche interactions. The diversifying phase persists — albeit with gradually slowing diversification — in the presence of general, nonspecific, fitness differences between strains, which break the assumption of tradeoffs inherent in much previous work. Building on a dynamical-mean field-theory analysis of the ecological dynamics, an approximate effective model captures the evolution of the diversity and distributions of key properties. This work establishes a potential scenario for understanding how the interplay between evolution and ecology — in particular coevolution of a bacterial and a generalist phage species — could give rise to the extensive fine-scale diversity that is ubiquitous in the microbial world. 

Rankings and scores are two common data types used by judges to express preferences and/or perceptions of quality in a collection of objects. Numerous models exist to study data of each type separately, but no unified statistical model captures both data types simultaneously without first performing data conversion. We propose the Mallows-Binomial model to close this gap, which combines a Mallows $$\phi$$ ranking model with Binomial score models through shared parameters that quantify object quality, a consensus ranking, and the level of consensus among judges. We propose an efficient tree-search algorithm to calculate the exact MLE of model parameters, study statistical properties of the model both analytically and through simulation, and apply our model to real data from an instance of grant panel review that collected both scores and partial rankings. Furthermore, we demonstrate how model outputs can be used to rank objects with confidence. The proposed model is shown to sensibly combine information from both scores and rankings to quantify object quality and measure consensus with appropriate levels of statistical uncertainty. 

Estimating excavation-induced ground surface displacements in urban areas is needed to assess potential structure damage. Empirical settlement distribution models have been widely used to estimate the zone of influence and ground response behind braced excavation walls. Three underground station excavations, part of the Los Angeles Metro’s K Line Crenshaw/LAX Transit Project, offer a unique opportunity to collect field instrumentation data to improve estimates of ground deformations. One excavation employed cross-lot braces and soldier piles and wood lagging while the other two were supported by cross-lot braces and stiffer Cutter-Soil-Mixing (CSM) walls. For the excavations with stiff support systems and relatively small wall movements, upward surface displacement or heave governed the ground surface response, while surface settlement was measured at the excavation with the more flexible wall system. This heave behavior is often masked by settlement caused by relatively large wall movements, and is thus commonly disregarded. By idealizing the excavation unloading as an upward strip load at the ground surface, the Boussinesq solution for elastic upward movement can be used in combination with a settlement component resulting from lateral wall movements to estimate the magnitude and distribution of excavation-induced surface displacements. 

Empirical methods for estimating tunneling-induced ground movements have been widely adopted in the tunneling industry. The transverse surface settlement profile can be described by a Gaussian curve or a modified Gaussian curve whose maximum value and trough width are related to volume loss. Volume loss in turn is related to soil type, tunnel geometry, and construction techniques. Several empirical equations have been developed based on the Gaussian curve and the assumptions of (1) trough width dependency on tunnel depth and ground condition; and (2) volume loss dependency on the ground type and construction techniques. For Earth Pressure Balance Machine (EPBM) tunneling, a volume loss of 0.5% in granular soils and 1%–2% the soft clay has been assumed in the past as an initial estimate. However, with complete filling and pressurization of both the shield (overcut) gap and the grouted tail gap around the lining, volume losses below 0.1% to 0.2% are being achieved in the alluvial granular and clay soils on current Los Angeles Metro tunneling projects. The LA Metro K Line Crenshaw/LAX transit project, tunneled from 2016 to 2018, has provided an opportunity to acquire and organize data on compatible data management systems, and evaluate the extensive field monitoring data for ground conditions specific to predominately granular soils in Old Alluvium. These data allow for the improvement of current empirical methods and correlations for predicting surface settlement induced by EPBM tunnels. The approximately 1-mi (1.6-km)-long, 20.6-ft (6.5-m)-diameter twin tunnels were excavated by an EPBM in a dense sand layer overlain by a silt/clay layer. The cover-to-diameter ratio was consistently about 2. The settlements and volume losses are observed to be heavily dependent on the face/shield pressures. In general, maintaining continuous pressures can significantly reduce settlements. An equation for estimating the volume loss based on the measured EPBM shield pressures is proposed. This equation can be used with the existing empirical methods to estimate the surface settlement profile transverse to the longitudinal axis of the tunnel. 

Conventional variational autoencoders fail in modeling correlations between data points due to their use of factorized priors. Amortized Gaussian process inference through GPVAEs has led to significant improvements in this regard, but is still inhibited by the intrinsic complexity of exact GP inference. We improve the scalability of these methods through principled sparse inference approaches. We propose a new scalable GPVAE model that outperforms existing approaches in terms of runtime and memory footprint, is easy to implement, and allows for joint end-to-end optimization of all components