Census block groups are used in location selection to determine the average drive time for all residents within a given radius to a proposed new store. The United States census uses 220,334 block groups, however the spatial distance between neighboring block groups in densely populated areas is small enough to cluster multiple block groups into a single unit. In this paper, we evaluate the efficiency and accuracy of drive time computations performed on clusters generated by our novel approach of constrained recursive reclustering as run on three traditional clustering algorithms-affinity propagation, k-means, and mean shift. We perform comparisons of our constrained recursive reclustering approach against drive times computed using the original census block group, and using clusters obtained by traditional reclustering. Unlike traditional clustering, where clustering is performed in a single pass, our approach continues reclustering each new cluster until a user specified stopping criteria is reached. We show that traditional clustering techniques generate sub-optimal clusters, with large spatial distances between the cluster centroid and cluster points making them unusable for computing drive times. Our approach provides reductions of 81.2%, 83.4%, and 10.2% for affinity propagation, k-means, and mean shift respectively when run on 220,334 census block groups. Using 200 randomly sampled locations each from Lowe's, CVS, and Walmart, we show that compared to the original block groups there is no statistically significant difference in drive time computations when using clusters generated by constrained recursive reclustering with affinity propagation for any of the three businesses, and with k-means for CVS and Walmart. While statistically significant differences are obtained with k-means for Lowe's and with mean shift for all three businesses, the differences are negligible, with the mean difference for each location set being within 30 s.

—We present an approach that performs automated detection of whether an older adult has performed cognitive tasks such as form filling or problem solving using RGB-D video data of older adults collected using the Microsoft Kinect v2 sensor. Our approach uses the variances of 25 joint points on the 3D skeleton obtained from the Kinect for training random forest classifiers to detect if cognitive tasks are performed, based on deviations in postural sway induced by cognitive tasks. We validate our approach using a dataset of 10 subjects performing the test on standing with eyes closed in the Berg Balance Scale (BBS) series of diagnostic tests before and after cognitive tasks. Using leave- one-subject-out cross-validation, we obtain an average detection accuracy of 69.5%, with accuracies of 60% and 79% at detecting that the test on standing with eyes closed was performed prior to and after cognitive tasks respectively. Our approach can be incorporated into intelligent health care systems to detect whether older adults have performed cognitively demanding activities that may induce stress or fatigue, and allow early intervention well before the occurrence of adverse events such as falls.

Location selection determines the feasibility of a new location by evaluating factors such as the drive time of customers, the number of potential customers, and the number and proximity of competitors to the new location. Traditional location selection approaches use census block group data to determine average customer drive times by computing the drive time from each block group to the proposed location and comparing it to all competitors within the area. However, since companies need to evaluate on the order of hundreds of thousands of potential locations and competitors, traditional location selection approaches prove to be computationally infeasible. In this paper we present an approach that generates an optimal set of clusters to speed up drive time calculations. Our approach is based on the insight that in urban areas block groups are comprised of a few adjacent city blocks, making the differences in drive times between neighboring block groups negligible. We use affinity propagation to initially cluster the census block groups. We use population and average distance between the cluster centroid and all points to recursively re-cluster the initial clusters. Our approach reduces the census data for the United States by 80% which provides a 5x speed when computing drive times. We sample 200 randomly generated locations across the United States and show that there is no statistically significant difference in the drive times when using the raw census data and our recursively clustered data. Additionally, for further validation we select 300 random Walmart stores across the United States and show that there is no statistically significant difference in the drive times.

Touch dynamics is a behavioral biometric that authenticates users by analyzing the characteristics of the touch gestures executed on mobile devices. Current research in this field has mostly focused on identifying the best algorithms and attributes to improve authentication performance. However, such systems must also be resilient against environmental variables. In this paper, we demonstrate that the user's posture, device size and configuration have a significant impact on the performance of touch-based authentication systems. Our results indicate that authentication accuracy increases with the device size. Furthermore, we conclude that using a device's 3-D orientation is necessary to attain better authentication performance. Our findings indicate that the features used in state-of-the-art touch-based authentication systems are insufficient to provide constant, reliable performance when any of the studied environmental variables change. With this paper, we release a new data set. Unlike the currently publicly available touch-based authentication datasets, our collection protocols control for all the studied variables. Our research study demonstrates threats to validity that noisy environmental conditions introduce to these currently available public datasets. This work is an extension of a previous publication. Presented user authentication approaches are unique and may have immediate benefits to the development of better touch-based authentication systems. (C) 2018 Published by Elsevier Inc.

Context Evaluating the adequacy of created wetlands to replace the functions of lost natural wetlands is important because wetland mitigation is a major tool used to offset wetland losses. However, measurements such as vegetative cover and presence of wildlife may not provide sufficient evidence that created wetlands are functioning properly. Thus, examining the ecology of wetland biota such as that of amphibians may be a more useful surrogate for function. Aims The objectives of this study were to compare the abundance of amphibian metamorphs and survival and growth of larval amphibians in created wetlands, relative to natural wetlands. Methods Amphibian metamorphs were trapped in created and natural wetlands during the spring (April–May) and summer (June–August) of 2009 and 2010, and 165 green frog (Lithobates clamitans) larvae were raised during the spring of 2010 in laboratory aquaria containing water from created or natural wetlands. Key results Abundance of spring peeper (Pseudacris crucifer) metamorphs decreased significantly from 2009 to 2010 and abundance of green frog metamorphs increased with habitat complexity, but both were unaffected by wetland type. Detection probability of metamorphs of both species was low, increased with water temperature and declined with month of observation. Survival, growth curves and mass were similar among green frog larvae raised in created and natural wetland aquaria. Conclusions Our results suggest that the created and natural wetlands we examined function similarly with respect to providing adequate breeding habitat for green frogs and spring peepers. Implications Wetlands created to offset the loss of natural wetlands, although generally not designed for the purpose of wildlife habitat, can function as adequate breeding habitat for generalist amphibians such as green frogs and spring peepers.

Context: Bug tracking systems play an important role in software maintenance. They allow both developers and users to submit problem reports on observed failures. However, by allowing anyone to submit problem reports, it is likely that more than one reporter will report on the same issue. Research in open source repositories has focused on two broad areas: determining the original report associated with each known duplicate, and assigning a developer to fix a particular problem.Objective: Limited research has been done in developing a fully automated triager, one that can first ascertain if a problem report is original or duplicate, and then provide a list of 20 potential matches for a duplicate report. We address this limitation by developing an automated triaging system that can be used to assist human triagers in bug tracking systems.Method: Our automated triaging system automatically assigns a label of original or duplicate to each incoming problem report, and provides a list of 20 suggestions for reports classified as duplicate. The system uses 24 document similarity measures and associated summary statistics, along with a suite of document property and user metrics. We perform our research on a lifetime of problem reports from the Eclipse, Firefox and Open Office repositories. Results: Our system can be used as a filtration aide, with high original recall exceeding 95% and low duplicate recall, or as a triaging guide, with balanced recall of approximately 70% for both originals and duplicates. Furthermore, the system reduces the workload on the triager by over 90%.Conclusions: Our work represents the first full scale effort at automatically triaging problem reports in open source repositories. By utilizing multiple similarity measures, we reduce the potential of false matches caused by the diversity of human language. (C) 2016 Elsevier B.V. All rights reserved.

Touch dynamics is a behavioral biometric that authenticates users by analyzing the characteristics of the touch gestures they execute on devices such as tablets and smartphones. The current research in this field has focused on identifying the best algorithms and the most effective attributes to improve authentication performance. However, a robust touch dynamics based authentication system for mobile devices must also be resilient against environmental variables such as user posture, movement, device size, device manufacturer, etc. In this work, we focus on two critical environmental variables that affect touch based authentication systems. We demonstrate that the user's posture and device size have a significant impact on the performance of touch based authentication systems. Our results indicate that authentication performance is proportional to the device size. Furthermore, we conclude that using a device's 3-D orientation is necessary to attain better authentication performance. Our findings indicate that the features used in state-of-the-art touch-based authentication systems are insufficient to provide constant, reliable performance when either the device size or user posture change. The effect of environmental variables on touch dynamics has not been explored. The results presented in this work are the first of its kind and important in the development of robust touch-based authentication systems. This study has immediate, applicable benefits to develop better authentication approaches touch dynamics.

Bug tracking systems play an important role in the development and maintenance of large-scale software systems. Having access to open source bug tracking systems has allowed researchers to take advantage of rich datasets and propose solutions to manage duplicate report classification, developer assignment and quality assessment. In spite of research advances, our understanding of the content of these repositories remains limited, primarily because of their size. In many cases, researchers analyze small portions of datasets thus limiting the understanding of the dynamics of problem reporting. The objective of this study is to explore the properties of two large-scale open source problem report repositories. The Eclipse dataset, at the time of download, consisted of 363; 770 reports spanning 11+ years, whereas Mozilla contained 699; 085 reports spanning 14+ years.Our research examines the evolution of datasets over time by analyzing the changes in the repository and the profiles of users who submit problem reports. We provide quantitative evidence on how submitter's maturity reduces the propensity to submit poor quality, insignificant or duplicate reports. We show that a diverse user base, characteristic of Mozilla, creates challenges for the development team as they spend more time triaging, rather than fixing, issues. Finally, we provide the research community with a series of observations and suggestions on how to study large-scale problem repositories.

The GlobalSoilMap project aims to create a global grid of a variety of soil functional properties at a fine resolution. Uncertainty surrounding these property estimates is of utmost importance when utilizing soil maps for predictive purposes. For the initial version of the map being produced of the United States, property values were estimated from the U.S. General Soil Map (STATSGO2) database, which is a broad-based inventory of soil data recorded across the United States. Multiple soil and non-soil (water, rock outcrop, urban areas, etc.) components are aggregated in STATSGO2 into groups of similar polygons, called map units. Each soil component includes various estimated property values for each horizon associated with that specific soil (if applicable), including an estimated upper, lower, and representative value. Hierarchically, STATSGO2 includes multiple horizon values for each component, which make up the map units. Previous work has used area-weighted means of the representative values for each component to develop a representative value for the map unit. For this study, prediction intervals were developed from the low and high estimated property values provided in the database. Instead of calculating a weighted mean of the low and high estimated property values, the lower prediction limit was determined as the lowest of the low values associated with any of the components in the map unit. The upper prediction limit was determined as the highest of the high values in the same manner. For each map unit, this method provided a unique prediction interval which was likely to encompass property values of soil typically found in that map unit. We empirically evaluated the soil property prediction intervals derived from STATSGO2 for three soil properties: organic carbon content, pH, and clay content. Using measured property data from up to 722 pedons from the National Cooperative Soil Survey database, prediction intervals were assessed by modeling their coverage accuracy over a set of external validation data. The effects of soil depth, soil order, temperature regime, and moisture regime on prediction interval coverage were analyzed, and coverage was found to be 87.6% for organic carbon, 90.6% for pH, and 86.4% for clay. It is shown that legacy data from the United States that includes low and high property methods can be used to represent uncertainty in the form of prediction intervals. Coverage based on these methods is only slightly below the nominal level of 95%. Consistency of these intervals was demonstrated across a variety of soil orders, temperature regimes, and moisture regimes.

Issue tracking systems play a critical role in software maintenance by allowing users and developers to submit problem reports for observed failures. A major problem in these systems is that two or more users can, and do, submit reports describing the same issue. Automated classification of such duplicate problem reports is an area of active research. The corpus of existing research shows a slow improvement in classification accuracy using relatively small subsets of problem report data. When applied to an entire project's problem repository, they exhibit a reduction in performance. In this paper we propose a novel duplicate report detection approach using multi-label classification. We use a suite of 24 duplicate classification techniques and MULAN software package to train a multi-label classifier. This multi-label classifier selects a set of similarity measures (from a pool of measures) that are most likely to find the true primary report. To demonstrate its effectiveness the method was tested on the entire Firefox repository. This data set encompasses 12+ years of problem reports and contains over 30,000 duplicate reports. Our results indicate that multi-label classification boosts the performance of the individual measures by up to 40% while returning overall results that match or outperform existing methods. The proposed method uses less than 1% of the dataset for training.