Spurious Observations Research Articles

DETECTING OUTLIER IN THE MULTIVARIATE DISTRIBUTION USING PRINCIPAL COMPONENTS

It is crucial to make inference out of the data at hand. It makes sense to discard spurious observations prior to application of statistical analysis. This study advances a procedure of determining outliers based from the principal components of the original variables. These variables are sorted and given weights based on the magnitude of their inner product with the principal components formulated from the centered and scaled variables. The weights are the corresponding variances explained by the principal components. The measure of proximity among observations is proportionate to the variance (eigenvalues) associated with the principal components. The methodology defines two distinct subintervals where the suspected outliers settle in one of these subintervals based on the proximity measures δo. On the merit of simulated data, the procedure detected 100 percent when the outliers are coming from distinct distribution. On the other hand, the procedure detected 98.7 per cent when the distribution of outliers have equal variance-covariance matrix with the outlier-free distribution and a slight difference in the vector of means.

Research on constrained localization of ultrasound geometric distribution based on FHN neurons.

The autopilot positioning process is mainly affected by three aspects: the first is the spatial geometric distribution of positioning sensors; the second is the screening of spurious observations; and the third is the equivalent ranging error. A constrained positioning method based on the geometric distribution of FitzHugh-Nagumo (FHN) neurons is proposed. To reduce the geometric accuracy factor, a Horizontal Dilution Of Precision value algorithm with a weight factor was proposed by considering the spatial geometric distribution of base stations and the geometric relationship of anchor points. This paper proposes a geometric constraint data processing method for the error of the pseudo-observation value. Finally, considering the significant weak signal perception ability of the biological nervous system, and the stochastic resonance phenomenon caused by noise can enhance the ability of the neuronal system to detect weak signals, an ultrasonic receiving method based on the stochastic resonance characteristics of FHN neuronal system is proposed, to enhance the signal and reduce noise. The results show that under the optimized base station layout and data geometric constraint processing, the ultrasonic wave based on FHN neuron improves the accuracy of spurious observations, reduces the calculation amount of geometric constraint processing, and reduces the positioning error by 66.67%, which provides a new direction for improving the positioning accuracy.

85 Automated Walk-Over Weighing System: Methods to Track Daily Body Mass and Growth in Grazing Steers

Abstract Body weight (BW) is a critical component for monitoring animal weight gain, body condition, nutritional status. Remote animal weighing systems facilitate frequent collection of animal BW, however, datasets often contain spurious data. The objective of this study was to describe the utility of using a remote Walk-over-Weigh system and subsequent methods for data cleaning. Beef steers (n = 10) were tagged with Electronic RFID tags (EID) in an improved pasture (~12.1 hectares) containing Bermuda and Tall Fescue and inter-seeded with Annual Ryegrass and grazed from Feb. – Dec. 2020. Static chute weights (n = 80) were collected monthly, and a WOW system placed by the water to remotely collect BW (n = 5,466). Data were first loaded into Program R and scanned for spurious data using each of 2 primary approaches, 1) the whole herd and individual means ± 1 standard deviation (SD) calculated daily or over the entire trial and 2) each of 3 data smoothing algorithms, which included a quadratic growth model, cubic splines, and polynomial regression. Then, data with spurious observations removed were paired with static chute weights and fitted to a linear model to measure accuracy (mean bias) and precision (R2) of WOW data. Whole herd mean ± 1SD and individual daily mean ± 1SD identified 1,204 and 1,516 spurious data, with mean bias of -12.46 and -15.37 KG and R2 of 0.90 and 0.68, respectively. Smoothing functions identified 1,707, 4,684, and 4,776 spurious points, with a mean bias of 13.61, -19.78, and 12.58 KG, and R2 of 0.94, 0.70, and 0.87 for quadratic growth models, cubic splines, and polynomial regression, respectively. These results indicate the utility of using a simple WOW system to collect data for measuring growth curves and using weight data in a real-time fashion to make management and marketing decisions.

Almost aggregations in the gravitational clustering to perform anomaly detection

Anomaly detection is the process of identifying observations that differ significantly from the norm in a dataset. Since there is not a proper formal definition of anomaly, different algorithms have arised to cope with the different variations of this idea, like novelty detection or outlier detection. Some of these algorithms have traditionally relied on prior knowledge on the data domain, or some degree of supervised learning in order to detect the irregular samples. In this work we propose a simulation based algorithm to detect those observations, based on gravitational forces, that requires not prior knowledge nor data labels to identify spurious observations. We do so by studying different generalizations of the aggregation of gravitational forces, and the resulting clusters obtained when the particles attract each other. We also compare our algorithm with other unsupervised anomaly detection algorithms, obtaining favourable results to our proposal.

Continuum dynamics and statistical correction of compositional heterogeneity in multivalent IDP oligomers resolved by single-particle EM

Multivalent intrinsically disordered protein (IDP) complexes are prevalent in biology and act in regulation of diverse processes, including transcription, signaling events, and the assembly and disassembly of complex macromolecular architectures. These systems pose significant challenges to structural investigation, due to continuum dynamics imparted by the IDP and compositional heterogeneity resulting from characteristic low-affinity interactions. Here, we developed a modular pipeline for automated single-particle electron microscopy (EM) distribution analysis of common but relatively understudied semi-ordered systems: ‘beads-on-a-string’ assemblies, composed of IDPs bound at multivalent sites to the ubiquitous ∼20 kDa cross-linking hub protein LC8. This approach quantifies conformational geometries and compositional heterogeneity on a single-particle basis, and statistically corrects spurious observations arising from random proximity of bound and unbound LC8. The statistical correction is generically applicable to oligomer characterization and not specific to our pipeline. Following validation, the approach was applied to the nuclear pore IDP Nup159 and the transcription factor ASCIZ. This analysis unveiled significant compositional and conformational diversity in both systems that could not be obtained from ensemble single particle EM class-averaging strategies, and new insights for exploring how these architectural properties might contribute to their physiological roles in supramolecular assembly and transcriptional regulation. We expect that this approach may be adopted to many other intrinsically disordered systems that have evaded traditional methods of structural characterization.

Inhabiting a Moth-Eaten Method

Inhabiting a Moth-Eaten Method Anila Gill (bio) The Punjab Civil Secretariat Library is housed in one of the many narrow back alleys near the Walled City in Lahore. It contains governmental records dating from the colonial era to the contemporary moment. Though the records pre-date the Partition of India and Pakistani Independence, they can only be accessed with Pakistani governmental approval: the archive, too, remains bound by state borders. From previous archival work in the region, I knew that material regarding the origins of the 1947 Partition form a fragmented, dismembered corpus. At the Punjab Civil Secretariat Library, I asked a librarian for all records pre-dating 1947. Some records I had seen before—the Indian Cinematograph Committee report from 1918 and its resulting Indian Cinematograph Act1—but the regional reports about cinematic regulation from the region of Punjab were new archival treasures that filled many of the historical gaps that had occasioned my research. Amongst these findings were ethnographic surveys from 1927 about the Punjab region that evince the colonial desire to know, categorize, and conquer the Indian subject. These surveys were the most decayed materials amongst the stack of records that the librarian laid down in front of me, but their decomposition was not the result of overuse or the erosive touch of human hands. Rather, insects had bored through each bound copy, cover to cover, leaving tunnels that appear at first as punctuation. These small creatures inhabit the archive while contributing to its destruction, modeling an anti-colonial orientation to research methodology. Their pestilent presence in the archive corrodes the justifications for colonial power affirmed in records from the colonial era; the historian must follow the [End Page 77] example of these moths to live in, pierce through, and inevitably destroy the colonial structures still managed and regulated by the postcolonial state. As the content of these ethnographic records consecrate and justify the colonial project, the holes puncturing these documents offer a form that configures an anti-colonial research method by centralizing absence and accounting for loss—a method that peers through the immaterial holes of the archive while looking askance at the material bounding it. A moth-eaten method is also a matter of perspective. The ethnographic survey identifies a man from the Jat tribe, native to the Punjab region, as "the most impatient of tribal or communal control, and the one whom asserts the freedom of the individual most strongly […] a Jat is a man who does what seems right in his own eyes, and sometimes what seems wrong also, and will not be said nay by any man […] He is independent, and he is self-willed, but he is reasonable, peaceably inclined if left alone, and not difficult to manage."2 With these spurious observations, the survey unveils its ultimate purpose: to characterize cultures that would comply with, and ultimately facilitate, the exigencies of the local colonial administration's regimes of control. While nakedly deploying the process by which colonial ethnographers leveraged their discipline to further affirm the means of domination over such regional tribes, the survey also reveals the contradictions of its analysis. The Jat does what is right, except for the times when he does what is wrong; he is deeply committed to personal freedom, yet compliant to managerial oversight in nature; reasonable yet unpredictable; unique but knowable, i.e., just as complex and uncategorizable as any other culture. The survey shows its seams, the survey has holes in its logic. The colonial conceit is one that imagines an absent other in need of settler presence. These holes are immanent with the possibility of an anti-colonial research method as they describe the absences of displaced populations during the decolonial Partition era, and they embody the historical moment as a diffractive and delayed punctum rather a documented, material event.3 In a public radio broadcast,4 Muhammad Ali Jinnah famously described Pakistan as "moth-eaten" to account for the absences left by displaced populations. As these absences continue to play out in ongoing sectarian violence within the ethno-religious nation-states of India and Pakistan, moth-eateness as a methodological metaphor offers a perspective on border history that allows...

SA1D-CNN: A Separable and Attention Based Lightweight Sensor Fault Diagnosis Method for Solar Insecticidal Lamp Internet of Things

Sensor faults can produce abnormal and spurious observations in the solar insecticidal lamp Internet of Things (SIL-IoTs) system. Early detection and identification of the sensor node's abnormality are critical to ensure the SIL-IoTs system's reliability. In this study, we propose a lightweight separable 1D convolution neural network that can be implemented in SIL-IoTs nodes to identify sensor faults, reduce detecting delay, and decrease data transmission. However, the reliability of data acquired by sensors is decreased because a SIL-IoTs node releases high voltage pulse discharge (a kind of electromagnetic interference) when pests collide with its metal mesh. This kind of data fluctuation impacts fault diagnosis accuracy. Consequently, fault-related feature maps and temporal signals are characterized via a novel time and channel attention module (TCAM) method, which contributes to separating electromagnetic interference noise from sensor faults of SIL-IoTs nodes. A real-world testbed is applied to validate the effectiveness of the proposed method on sensor fault diagnosis in the SIL-IoTs system. Experimental results demonstrate that the proposed method can detect four typical sensor faults with the best trade-off between accuracy (99.9% average accuracy and 97.6% average F1-score) and efficiency (351KB inference model size and 4.33W average energy consumption on Raspberry Pi).

Robust parameter estimation based on the generalized log-likelihood in the context of Sharma-Taneja-Mittal measure.

The problem of obtaining physical parameters that cannot be directly measured from observed data arises in several scientific fields. In the classic approach, the well-known maximum likelihood estimation associated with a Gaussian distribution is employed to obtain the model parameters of a complex system. Although this approach is quite popular in statistical physics, only a handful of spurious observations (outliers) make this approach ineffective, violating the Gauss-Markov theorem. In this work, starting from the generalized logarithmic function associated to the Sharma-Taneja-Mittal (STM) information measure, we propose an outlier-resistant approach based on the generalized log-likelihood estimation. In particular, our proposal deforms the Gaussian distribution based on a two-parameter generalization of the ordinary logarithmic function. We have tested the effectiveness of our proposal considering a classic geophysical inverse problem with a very noisy data set. The results show that the task of obtaining physical parameters based on the STM measure from noisy data with several outliers outperforms the classic approach, and therefore, our proposal is a useful tool for statistical physics, information theory, and statistical inference problems.

A Flow‐Dependent Targeted Observation Method for Ensemble Kalman Filter Assimilation Systems

AbstractIn this study, we developed a flow‐dependent sequential assimilation‐based targeted observation method by minimizing the analysis error variance under the framework of the ensemble Kalman filter (EnKF). This approach considers the flow‐dependent variation in background error statistics when identifying optimal observational sites through the sequential assimilation method. Covariance localization is also introduced in this method, enabling computational efficiency and eliminating impacts from spurious observations. By quantifying the reduction in analysis error variances, the proposed method could estimate the potential improvements by each optimal observation while assimilated. With this method, we design an optimal observational array for sea level anomaly (SLA) prediction in the tropical Indian Ocean (TIO), which is implemented using a fully coupled climate model, the Community Earth System Model (CESM), in conjunction with a coupled assimilation system. The optimal observational array detected from this method was found to theoretically reduce the initial uncertainty by up to approximately 60% of the error variance. An observing system simulation experiment (OSSE) using the CESM and the coupled assimilation system, which was designed for validation purposes, confirms the theoretical reduction in the analysis error variance by the optimal observation array.

A motion strategy for exploration driven by an automaton activating feedback-based controllers

This paper addresses the problem of exploring an unknown, planar, polygonal and simply connected environment. To explore the environment, the robot follows the environment boundary. In the first part of this paper, we propose a motion policy based on simple sensor feedback and a complete exploration strategy is represented as a Moore machine. The proposed motion policy is based on the paradigm of avoiding the state estimation; there is a direct mapping from observation to control. We present the theoretical conditions guaranteeing that the robot discovers the largest possible region of the environment. In the second part of the paper, we propose an automaton that filters spurious observations to activate feedback-based controllers. We propose a practical control scheme whose objective is to maintain a desired distance between the robot and the boundary of the environment. The approach is able to deal with imprecise robot’s observations and controls, and to take into account variations in the robot’s velocities. The control scheme switches controllers according to observations obtained from the robots sensor. Our control scheme aims to maintain the continuity of angular and linear velocities of the robot in spite of the switching between controllers. All the proposed techniques have been implemented and both simulations and experiments in a real robot are presented.

Sharing refuges on arid islands: ecological and social influence on aggregation behaviour of wall geckos.

BackgroundThe extent of social behaviour among reptiles is underappreciated. Two types of aggregations are recognized in lizards: ecological and social, i.e., related to the attraction to a site or to animals of the same species, respectively. As most lizards are territorial, aggregations increase the probability of aggressive interactions among individuals, a density-dependent behaviour.MethodsAfter some spurious observations of aggregation behaviour in the endemic Cabo Verde nocturnal gecko Tarentola substituta, we conducted a field-based study in order to thoroughly characterize it. We sampled 48 transects and 40 10 × 10 m quadrats on São Vicente Island to describe the incidence, size and composition of aggregations and to study the effect of gecko and refuge density, plus refuge quality, on refuge sharing. We hypothesize that when density of animals and scarcity of high-quality refuges is higher, lizards have increased probability of aggregating. We also predict a consistent pattern of size and composition of groups (male–female pairs, only one adult male per group) throughout the year if there is a selected behaviour to avoid agonistic interactions, and low thermal advantage to aggregating individuals.ResultsWe present one of the first evidences of aggregation for Phyllodactylidae geckos. We found that T. substituta forms aggregations around 30–40% of the time, and that refuges are almost always shared by a female-male pair, sometimes with a juvenile, probably a mechanism to avoid aggressive interactions. We also observed that refuge sharing is dependent on refuge quality, as medium–large (thermally more stable and positively selected) rocks are shared much more frequently than small ones, but independent of adult sizes. Refuge sharing is also directly related to the density of geckos and inversely related to the density of high-quality refuges. We found no relation between body temperatures of geckos and refuge sharing when controlling the effect of rock/air temperature, suggesting that huddling does not improve thermoregulation.DiscussionOur results suggest that in this harsh environment (rocks reach 46 °C) aggregation incidence is mainly driven by an ecological factor (scarcity of high-quality refuges) and its intersexual composition by social factors (avoidance of agonistic interactions by males, and possible increased reproductive success of the pair). This study sheds some light on the little explored gecko aggregation behaviour and other studies should follow.

Assessing Low-Intensity Relationships in Complex Networks.

Many large network data sets are noisy and contain links representing low-intensity relationships that are difficult to differentiate from random interactions. This is especially relevant for high-throughput data from systems biology, large-scale ecological data, but also for Web 2.0 data on human interactions. In these networks with missing and spurious links, it is possible to refine the data based on the principle of structural similarity, which assesses the shared neighborhood of two nodes. By using similarity measures to globally rank all possible links and choosing the top-ranked pairs, true links can be validated, missing links inferred, and spurious observations removed. While many similarity measures have been proposed to this end, there is no general consensus on which one to use. In this article, we first contribute a set of benchmarks for complex networks from three different settings (e-commerce, systems biology, and social networks) and thus enable a quantitative performance analysis of classic node similarity measures. Based on this, we then propose a new methodology for link assessment called z* that assesses the statistical significance of the number of their common neighbors by comparison with the expected value in a suitably chosen random graph model and which is a consistently top-performing algorithm for all benchmarks. In addition to a global ranking of links, we also use this method to identify the most similar neighbors of each single node in a local ranking, thereby showing the versatility of the method in two distinct scenarios and augmenting its applicability. Finally, we perform an exploratory analysis on an oceanographic plankton data set and find that the distribution of microbes follows similar biogeographic rules as those of macroorganisms, a result that rejects the global dispersal hypothesis for microbes.

ICESat‐derived inland water surface spot heights

AbstractAccurate measurement of water surface height is key to many fields in hydrology and limnology. Satellite radar and laser altimetry have been shown to be useful means of obtaining such data where no ground gauging stations exist, and the accuracy of different satellite instruments is now reasonably well understood. Past validation studies have shown water surface height data from the ICESat instrument to have the highest vertical accuracy (mean absolute errors of ∼10 cm for ICESat, compared, for example, with ∼28 cm from Envisat), yet no freely available source of processed ICESat data currently exists for inland water bodies. Here we present a database of processed and quality checked ICESat‐derived inland water surface heights (IWSH) for water bodies greater than 3 arc sec (∼92 m at the equator) in width. Four automated methods for removing spurious observations or outliers were investigated, along with the impact of using different water masks. We find that the best performing method ensures that observations used are completely surrounded by water in the SRTM Water Body data. Using this method for removing spurious observations, we estimate transect‐averaged water surface heights at 587,292 unique locations from 2003 to 2009, with the number of locations proportional to the size of the river.

The flush early and avoid the rush hypothesis holds after accounting for spontaneous behavior

When approached by a predator, prey make economic decisions between remaining where they are and obtaining benefits from their current activity or leaving and enhancing their safety. The “flush early and avoid the rush” hypothesis suggests that animals that flee to escape approaching threats flee soon after they become alert to an approaching predator so as to reduce any costs incurred by ongoing monitoring of the predator. This hypothesis has been supported by several studies, but some researchers argue the relationship may be partially or entirely a consequence of bouts of spontaneous vigilance and/or bouts of spontaneous locomotion (vigilance or locomotion that occur when the animal is unaffected by a predator), rather than an economic decision related to the approaching predator. If this were true, spontaneous vigilance might incorrectly be recorded as alert distance (predator–prey distance when the prey becomes aware of and begins to monitor the predator) and spontaneous locomotion might be incorrectly recorded as flight initiation distance (predator–prey distance when escape begins). To evaluate these potential effects, we recorded the intervals between bouts of spontaneous vigilance and locomotion by yellow-bellied marmots (Marmota flaviventris). We used these baseline rates to conduct a series of alert distance–flight initiation distance regressions after removing potentially spurious observations recorded as alert distance or flight initiation distance. Although spontaneous vigilance and spontaneous locomotion may lead to artifactual increases in flight initiation distance as alert distance increases, the fundamental relationship remains after effects of spontaneous movements have been removed, supporting the flush early and avoid the rush hypothesis. We tested a key challenge of the “flush early and avoid the rush” (FEAR) hypothesis; our results provide strong support for the hypothesis.

Recursive RANSAC: Multiple Signal Estimation with Outliers

The random sample consensus (RANSAC) algorithm is frequently used in computer vision to estimate the parameters of a signal in the presence of noisy and even spurious observations called gross errors. Instead of just one signal, we desire to estimate the parameters of multiple signals, where at each time step a set of observations of generated from the underlying signals and gross errors are received. In this paper, we develop the recursive RANSAC (RRANSAC) algorithm to solve the inherent data association problem and recursively estimate the parameters of multiple signals without prior knowledge of the number of true signals. We compare the performance of RRANSAC with several existing algorithms, and also demonstrate the capabilities of RRANSAC in an aerial geolocation problem.

You are what you eat: dietary salt intake and renin–angiotensin blockade in diabetic nephropathy

Interactions between sodium intake, the renin angiotensin system and renal and cardiovascular outcomes are incompletely understood. Lambers Heerspink et al's analysis shows that angiotensin receptor blockade (ARB) improves diabetic nephropathy and cardiovascular disease more when dietary sodium intake is low and suggests possible harm when sodium intake is high. These findings highlight dietary salt as a modifiable cardiovascular and renal risk factor and emphasize the need for need for detailed mechanistic studies.

Vegetation height and cover fraction between 60° S and 60° N from ICESat GLAS data

Abstract. We present new coarse resolution (0.5° × 0.5°) vegetation height and vegetation-cover fraction data sets between 60° S and 60° N for use in climate models and ecological models. The data sets are derived from 2003–2009 measurements collected by the Geoscience Laser Altimeter System (GLAS) on the Ice, Cloud and land Elevation Satellite (ICESat), the only LiDAR instrument that provides close to global coverage. Initial vegetation height is calculated from GLAS data using a development of the model of Rosette et al. (2008) with with further calibration on desert sites. Filters are developed to identify and eliminate spurious observations in the GLAS data, e.g. data that are affected by clouds, atmosphere and terrain and as such result in erroneous estimates of vegetation height or vegetation cover. Filtered GLAS vegetation height estimates are aggregated in histograms from 0 to 70 m in 0.5 m intervals for each 0.5° × 0.5°. The GLAS vegetation height product is evaluated in four ways. Firstly, the Vegetation height data and data filters are evaluated using aircraft LiDAR measurements of the same for ten sites in the Americas, Europe, and Australia. Application of filters to the GLAS vegetation height estimates increases the correlation with aircraft data from r = 0.33 to r = 0.78, decreases the root-mean-square error by a factor 3 to about 6 m (RMSE) or 4.5 m (68% error distribution) and decreases the bias from 5.7 m to −1.3 m. Secondly, the global aggregated GLAS vegetation height product is tested for sensitivity towards the choice of data quality filters; areas with frequent cloud cover and areas with steep terrain are the most sensitive to the choice of thresholds for the filters. The changes in height estimates by applying different filters are, for the main part, smaller than the overall uncertainty of 4.5–6 m established from the site measurements. Thirdly, the GLAS global vegetation height product is compared with a global vegetation height product typically used in a climate model, a recent global tree height product, and a vegetation greenness product and is shown to produce realistic estimates of vegetation height. Finally, the GLAS bare soil cover fraction is compared globally with the MODIS bare soil fraction (r = 0.65) and with bare soil cover fraction estimates derived from AVHRR NDVI data (r = 0.67); the GLAS tree-cover fraction is compared with the MODIS tree-cover fraction (r = 0.79). The evaluation indicates that filters applied to the GLAS data are conservative and eliminate a large proportion of spurious data, while only in a minority of cases at the cost of removing reliable data as well. The new GLAS vegetation height product appears more realistic than previous data sets used in climate models and ecological models and hence should significantly improve simulations that involve the land surface.

Cluster observations of trapped ions interacting with magnetosheath mirror modes

Abstract. Mirror modes are among the most intense low frequency plasma wave phenomena observed in the magnetosheaths of magnetized planets. They appear as large amplitude non-propagating fluctuations in the magnetic field magnitude and plasma density. These structures are widely accepted to represent a non-linear stage of the mirror instability, dominant in plasmas with large ion beta and a significant ion temperature anisotropy T⊥/T∥>1. It has long been recognized that the mirror instability both in the linear and non-linear stage is a kinetic process and that the behavior of resonant particles at small parallel velocities is crucial for its development and saturation. While the dynamics of the instability and the effect of trapped particles have been studied extensively in theoretical models and numerical simulations, only spurious observations of the trapped ions were published to date. In this work we used data from the Cluster spacecraft to perform the first detailed experimental study of ion velocity distribution associated with mirror mode oscillations. We show a conclusive evidence for the predicted cooling of resonant ions at small parallel velocities and heating of trapped ions at intermediate pitch angles.

Methods and Limitations of Assessing New Noninvasive Tests

Outcomes-based approaches are the preferred methodology for technology validation. As discussed in part I of this review, the difficulties of performing an unbiased diagnostic evaluation are increasingly appreciated.1 An outcomes-based approach is advantageous in that it mimics the clinical application of testing. By risk stratifying patients, the results of this approach can be applied directly to clinical practice.2 Nonetheless, outcomes-based technology validation is not without challenges and limitations. These issues include the need for multivariable modeling for observational data, end point selection, and limitations of estimating posttest risk. Finally, it is increasingly appreciated that the future “gold standard” for outcomes-based assessments will be demonstrating whether imaging can identify which therapeutic approach optimizes patient benefit rather than merely identifying risk. ### Study Design Requirements of imaging studies include a relevant study population, comparison with an appropriate control group, and follow-up for outcomes. Designing randomized controlled trials (RCTs) that address imaging questions is challenging. The RCT may utilize imaging results as inclusion criteria or the basis for therapeutic assignment. Randomization to strategies with versus without imaging is problematic. The use of imaging per se affects outcomes only if a therapy is triggered; hence, in these studies, therapy and imaging results must be linked.3,4 Comparisons of imaging methods/modalities must mandate that test results are acted on rather than “available to the physician” (thus, an emphasis on efficacy rather than effectiveness). Because observational studies are far more common, they will be our focus. Although limited by inherent design flaws (eg, selection biases, potentially spurious observations, missing covariates), patients in observational studies better represent those seen in practice. #### Patient Selection Unlike diagnostic-based validation in which only patients referred to a gold standard after testing are included, in prognostic-based approaches all eligible patients are followed up at a preselected time point after testing to determine their status relative …

A Method for Judicious Fusion of Inconsistent Multiple Sensor Data

One of the major problems in sensor fusion is that sensors frequently provide spurious observations which are difficult to predict and model. The spurious measurements from sensors must be identified and eliminated since their incorporation in the fusion pool might lead to inaccurate estimation. This paper presents a unified sensor fusion strategy based on a modified Bayesian approach that can automatically identify the inconsistency in sensor measurements so that the spurious measurements can be eliminated from the data fusion process. The proposed method adds a term to the commonly used Bayesian formulation. This term is an estimate of the probability that the data is not spurious, based upon the measured data and the unknown value of the true state. In fusing two measurements, it has the effect of increasing the variance of the posterior distribution when measurement from one of the sensors is inconsistent with respect to the other. The increase or decrease in variance can be estimated using the information theoretic measure entropy. The proposed strategy was verified with the help of extensive computations performed on simulated data from three sensors. A comparison was made between two different fusion schemes: centralized fusion in which data obtained from all sensors were fused simultaneously, and a decentralized or sequential Bayesian scheme that proved useful for identifying and eliminating spurious data from the fusion process. The simulations verified that the proposed strategy was able to identify spurious sensor measurements and eliminate them from the fusion process, thus leading to a better overall estimate of the true state. The proposed strategy was also validated with the help of experiments performed using stereo vision cameras, one infrared proximity sensor, and one laser proximity sensor. The information from these three sensing sources was fused to obtain an occupancy profile of the robotic workspace