Adaptive two-stage inverse sampling design to estimate density, abundance, and occupancy of rare and clustered populations

Sampling rare and clustered populations is challenging because of the effort required to find rare units. Heuristically, a practitioner would prefer to discontinue sampling in areas where rare units of interest are apparently extremely sparse or absent. We take advantage of the characteristics of inverse sampling to adaptively inform practitioners when it is efficient to move on to sample new areas. We introduce Adaptive Two-stage Inverse Sampling (ATIS), which is designed to leave a selected area after observation of an a priori number of only non-rare units and to continue sampling in the area when rare units are observed. ATIS is efficient in many cases and yields more rare units than conventional sampling for a rare and clustered population. We derive unbiased estimators of population total and variance. We also introduce an easy-to-compute estimator, which is nearly as efficient as the unbiased estimator. A simulation study on a rare plant population of buttercups (Ranunculus) shows that ATIS even with the easy-to-compute estimator is more efficient than its conventional sampling counterparts and is more efficient than Two-stage Adaptive Cluster Sampling (TACS) for small and moderate final sample sizes. Additional simulations reveal that ATIS is efficient for binary data (e.g., presence or absence) whereas TACS is inefficient for binary data. The overall results indicate that ATIS is consistently efficient compared to conventional sampling and to adaptive cluster sampling in some important cases.

Controlling Fairness and Bias in Dynamic Learning-to-Rank (Extended Abstract)

Rankings are the primary interface through which many online platforms match users to items (e.g. news, products, music, video). In these two-sided markets, not only do the users draw utility from the rankings, but the rankings also determine the utility (e.g. exposure, revenue) for the item providers (e.g. publishers, sellers, artists, studios). It has already been noted that myopically optimizing utility to the users -- as done by virtually all learning-to-rank algorithms -- can be unfair to the item providers. We, therefore, present a learning-to-rank approach for explicitly enforcing merit-based fairness guarantees to groups of items (e.g. articles by the same publisher, tracks by the same artist). In particular, we propose a learning algorithm that ensures notions of amortized group fairness, while simultaneously learning the ranking function from implicit feedback data. The algorithm takes the form of a controller that integrates unbiased estimators for both fairness and utility, dynamically adapting both as more data becomes available. In addition to its rigorous theoretical foundation and convergence guarantees, we find empirically that the algorithm is highly practical and robust.

Two Different Classes of Shrinkage Estimators for the Scale Parameter of the Rayleigh Distribution

Shrinkage estimators are introduced for the scale parameter of the Rayleigh distribution by using two different shrinkage techniques. The mean squared error properties of the proposed estimator have been derived. The comparison of proposed classes of the estimators is made with the respective conventional unbiased estimators by means of mean squared error in the simulation study. Simulation results show that the proposed shrinkage estimators yield smaller mean squared error than the existence of unbiased estimators.

A Theory of Higher Order Interactions Between Sensitive Variables: Empirical Evidences and an Application to a Variety of Smoking

In carrying out surveys involving sensitive characteristics, randomized response models have been considered among the best techniques since they provide the maximum privacy protection to the respondents and procure honest responses. Over the years, researchers have carried out studies on the estimation of proportions of the population possessing sensitive characteristics. However, there is a paucity of research studies that have addressed higher order interactions between these sensitive characters. In this article, we develop a new theory based on three proposed randomized response models which we name as: simple model, semi-crossed model, and fully crossed model. Twenty-one new unbiased estimators of seven parameters are introduced, their variance expressions are derived, and unbiased estimators of variances are developed. The three models are compared under various values of the parameters by computing the percent relative efficiency of one model over another model. The most efficient model is then applied to study the population proportions of three varieties of smoking habits among students, and their first- and second-order interactions. The last four sections (Ninth to Twelfth) are verifications of theoretical results using the Cramer–Rao lower bounds of variances for the developed 21 new estimators in randomized response sampling.

Online Sampling of Temporal Networks

Temporal networks representing a stream of timestamped edges are seemingly ubiquitous in the real world. However, the massive size and continuous nature of these networks make them fundamentally challenging to analyze and leverage for descriptive and predictive modeling tasks. In this work, we propose a general framework for temporal network sampling with unbiased estimation. We develop online, single-pass sampling algorithms, and unbiased estimators for temporal network sampling. The proposed algorithms enable fast, accurate, and memory-efficient statistical estimation of temporal network patterns and properties. In addition, we propose a temporally decaying sampling algorithm with unbiased estimators for studying networks that evolve in continuous time, where the strength of links is a function of time, and the motif patterns are temporally weighted. In contrast to the prior notion of a △ t -temporal motif, the proposed formulation and algorithms for counting temporally weighted motifs are useful for forecasting tasks in networks such as predicting future links, or a future time-series variable of nodes and links. Finally, extensive experiments on a variety of temporal networks from different domains demonstrate the effectiveness of the proposed algorithms. A detailed ablation study is provided to understand the impact of the various components of the proposed framework.

Cardinality estimation for random stopping sets based on Poisson point processes

We construct unbiased estimators for the distribution of the number of vertices inside random stopping sets based on a Poisson point process. Our approach is based on moment identities for stopping sets, showing that the random count of points inside the complement of a stopping set S has a Poisson distribution conditionally to S. The proofs do not require the use of set-indexed martingales, and our estimators have a lower variance when compared to standard sampling. Numerical simulations are presented for examples such as the convex hull and the Voronoi flower of a Poisson point process, and their complements.

Properties and unbiased estimation of F- and D-statistics in samples containing related and inbred individuals

The Patterson F- and D-statistics are commonly-used measures for quantifying population relationships and for testing hypotheses about demographic history. These statistics make use of allele frequency information across populations to infer different aspects of population history, such as population structure and introgression events. Inclusion of related or inbred individuals can bias such statistics, which may often lead to the filtering of such individuals. Here we derive statistical properties of the F- and D-statistics, including their biases due to finite sample size or the inclusion of related or inbred individuals, their variances, and their corresponding mean squared errors. Moreover, for those statistics that are biased, we develop unbiased estimators and evaluate the variances of these new quantities. Comparisons of the new unbiased statistics to the originals demonstrates that our newly-derived statistics often have lower error across a wide population parameter space. Furthermore, we apply these unbiased estimators using several global human populations with the inclusion of related individuals to highlight their application on an empirical dataset. Finally, we implement these unbiased estimators in open-source software package funbiased for easy application by the scientific community.