Search

The Extended Crosswise Model is a popular randomized response design that employs a sensitive and a randomized innocuous statement, and asks respondents if one of these statements is true, or that none or both are true. The model has a degree of freedom to test for response biases, but is unable to detect random answering. In this paper, we propose two new methods to indirectly estimate and correct for random answering. One method uses a non-sensitive control statement and a quasi-randomized innocuous statement to which both answers are known to estimate the proportion of random respondents. The other method assigns less weight in the estimation procedure to respondents who complete the survey in an unrealistically short time. For four surveys among elite athletes, we use these methods to correct the prevalence estimates of doping use for random answering.

Multiple systems estimation uses samples that each cover part of a population to obtain a total population size estimate. Ideally, all the available samples are used, but if some samples are available (much) later, one may use only the samples that are available early. Under some regularity conditions, including sample independence, two samples is enough to obtain an asymptotically unbiased population size estimate. However, the assumption of sample independence may be unrealistic, especially when samples are derived from administrative sources. The sample independence assumption can be relaxed when three or more samples are used, which is therefore generally recommended. This may be a problem if the third sample is available much later than the first two samples. Therefore, in this paper we propose a new approach that deals with this issue by utilising older samples, using the so-called expectation maximisation algorithm. This leads to a population size nowcast estimate that is asymptotically unbiased under more relaxed assumptions than the estimate based on two samples. The resulting nowcasting model is applied to the problem of estimating the number of homeless people in The Netherlands, which leads to reasonably accurate nowcast estimates.

Popular word embedding methods such as GloVe and Word2Vec are related to the factorization of the pointwise mutual information (PMI) matrix. In this paper, we link correspondence analysis (CA) to the factorization of the PMI matrix. CA is a dimensionality reduction method that uses singular value decomposition (SVD), and we show that CA is mathematically close to the weighted factorization of the PMI matrix. In addition, we present variants of CA that turn out to be successful in the factorization of the word-context matrix, i.e. CA applied to a matrix where the entries undergo a square-root transformation (ROOT-CA) and a root-root transformation (ROOTROOT-CA). While this study focuses on traditional static word embedding methods, to extend the contribution of this paper, we also include a comparison of transformer-based encoder BERT, i.e. contextual word embedding, with these traditional methods. An empirical comparison among CA- and PMI-based methods as well as BERT shows that overall results of ROOT-CA and ROOTROOT-CA are slightly better than those of the PMI-based methods and are competitive with BERT.

Correspondence analysis (CA) is a popular technique to visualize the relationship between two categorical variables. CA uses the data from a two-way contingency table and is affected by the presence of outliers. The supplementary points method is a popular method to handle outliers. Its disadvantage is that the information from entire rows or columns is removed. However, outliers can be caused by cells only. In this paper, a reconstitution algorithm is introduced to cope with such cells. This algorithm can reduce the contribution of cells in CA instead of deleting entire rows or columns. Thus the remaining information in the row and column involved can be used in the analysis. The reconstitution algorithm is compared with two alternative methods for handling outliers, the supplementary points method and MacroPCA. It is shown that the proposed strategy works well.

Technology-Assisted Review (TAR) aims to reduce the human effort required for screening processes such as abstract screening for systematic literature reviews. Human reviewers label documents as relevant or irrelevant during this process, while the system incrementally updates a prediction model based on the reviewers' previous decisions. After each model update, the system proposes new documents it deems relevant, to prioritize relevant documentsover irrelevant ones. A stopping criterion is necessary to guide users in stopping the review process to minimize the number of missed relevant documents and the number of read irrelevant documents. In this paper, we propose and evaluate a new ensemble-based Active Learning strategy and a stopping criterion based on Chao's Population Size Estimator that estimates the prevalence of relevant documents in the dataset. Our simulation study demonstrates that this criterion performs well on several datasets and is compared to other methods presented in the literature.

If part of a population is hidden but two or more sources are available that each cover parts of this population, dual- or multiple-system(s) estimation can be applied to estimate this population. For this it is common to use the log-linear model, estimated with maximum likelihood. These maximum likelihood estimates are based on a non-linear model and therefore suffer from finite-sample bias, which can be substantial in case of small samples or a small population size. This problem was recognised by Chapman, who derived an estimator with good small sample properties in case of two available sources. However, he did not derive an estimator for more than two sources. We propose an estimator that is an extension of Chapman's estimator to three or more sources and compare this estimator with other bias-reduced estimators in a simulation study. The proposed estimator performs well, and much better than the other estimators. A real data example on homelessness in the Netherlands shows that our proposed model can make a substantial difference.

We present a method to generate contingency tables that follow loglinear models with prescribed marginal probabilities and dependence structures. We make use of (loglinear) Poisson regression, where the dependence structures, described using odds ratios, are implemented using an offset term. We apply this methodology to carry out simulation studies in the context of population size estimation using dual system and triple system estimators, popular in official statistics. These estimators use contingency tables that summarise the counts of elements enumerated or captured within lists that are linked. The simulation is used to investigate these estimators in the situation that the model assumptions are fulfilled, and the situation that the model assumptions are violated.

Both latent semantic analysis (LSA) and correspondence analysis (CA) are dimensionality reduction techniques that use singular value decomposition (SVD) for information retrieval. Theoretically, the results of LSA display both the association between documents and terms, and marginal effects; in comparison, CA only focuses on the associations between documents and terms. Marginal effects are usually not relevant for information retrieval, and therefore, from a theoretical perspective CA is more suitable for information retrieval. In this paper, we empirically compare LSA and CA. The elements of the raw document-term matrix are weighted, and the weighting exponent of singular values is adjusted to improve the performance of LSA. We explore whether these two weightings also improve the performance of CA. In addition, we compare the optimal singular value weighting exponents for LSA and CA to identify what the initial dimensions in LSA correspond to. The results for four empirical datasets show that CA always performs better than LSA. Weighting the elements of the raw data matrix can improve CA; however, it is data dependent and the improvement is small. Adjusting the singular value weighting exponent usually improves the performance of CA; however, the extent of the improved performance depends on the dataset and number of dimensions. In general, CA needs a larger singular value weighting exponent than LSA to obtain the optimal performance. This indicates that CA emphasizes initial dimensions more than LSA, and thus, margins play an important role in the initial dimensions in LSA.

Compositional data are non-negative data collected in a rectangular matrix with a constant row sum. Due to the non-negativity the focus is on conditional proportions that add up to 1 for each row. A row of conditional proportions is called an observed budget. Latent budget analysis (LBA) assumes a mixture of latent budgets that explains the observed budgets. LBA is usually fitted to a contingency table, where the rows are levels of one or more explanatory variables and the columns the levels of a response variable. In prospective studies, there is only knowledge about the explanatory variables of individuals and interest goes out to predicting the response variable. Thus, a form of LBA is needed that has the functionality of prediction. Previous studies proposed a constrained neural network (NN) extension of LBA that was hampered by an unsatisfying prediction ability. Here we propose LBA-NN, a feed forward NN model that yields a similar interpretation to LBA but equips LBA with a better ability of prediction. A stable and plausible interpretation of LBA-NN is obtained through the use of importance plots and table, that show the relative importance of all explanatory variables on the response variable. An LBA-NN-K- means approach that applies K-means clustering on the importance table is used to produce K clusters that are comparable to K latent budgets in LBA. Here we provide different experiments where LBA-NN is implemented and compared with LBA. In our analysis, LBA-NN outperforms LBA in prediction in terms of accuracy, specificity, recall and mean square error. We provide open-source software at GitHub.

Latent semantic analysis (LSA) and correspondence analysis (CA) are two techniques that use a singular value decomposition (SVD) for dimensionality reduction. LSA has been extensively used to obtain low-dimensional representations that capture relationships among documents and terms. In this article, we present a theoretical analysis and comparison of the two techniques in the context of document-term matrices. We show that CA has some attractive properties as compared to LSA, for instance that effects of margins, i.e. sums of row elements and column elements, arising from differing document-lengths and term-frequencies are effectively eliminated, so that the CA solution is optimally suited to focus on relationships among documents and terms. A unifying framework is proposed that includes both CA and LSA as special cases. We empirically compare CA to various LSA based methods on text categorization in English and authorship attribution on historical Dutch texts, and find that CA performs significantly better. We also apply CA to a long-standing question regarding the authorship of the Dutch national anthem Wilhelmus and provide further support that it can be attributed to the author Datheen, amongst several contenders.

We investigate use of two or more linked registers, or lists, for both population size estimation and to investigate the relationship between variables appearing on all or only some registers. This relationship is usually not fully known because some individuals appear in only some registers, and some are not in any register. These two problems have been solved simultaneously using the EM algorithm. We extend this approach to estimate the size of the indigenous M\=aori population in New Zealand, leading to several innovations: (1) the approach is extended to four registers (including the population census), where the reporting of M\=aori status differs between registers; (2) some individuals in one or more registers have missing ethnicity, and we adapt the approach to handle this additional missingness; (3) some registers cover subsets of the population by design. We discuss under which assumptions such structural undercoverage can be ignored and provide a general result; (4) we treat the M\=aori indicator in each register as a variable measured with error, and embed a latent class model in the multiple system estimation to estimate the population size of a latent variable, interpreted as the true M\=aori status. Finally, we discuss estimating the M\=aori population size from administrative data only. Supplementary materials for our article are available online., Comment: 34 pages, 4 figures, submitted to JASA

We read with interest the article by di Cecco et al. (2018), but have reservations about the usefulness of the latent class model specifically for estimating overcoverage. In particular, we question the interpretation of the parameters of the fitted latent class model., Comment: 5 pages

In this paper it is shown that under certain conditions there is a relationship between the parameter estimation of the Fellegi--Sunter probabilistic linkage model and dual system estimation. This relationship can be used as the basis of an approach to population size estimation. In this case it is sufficient to estimate the parameters of the linkage model in order to obtain the population size estimate. Neither classification of the record pairs into links and non-links, nor forcing the records into a series of 1-1 matches, nor clerical review of the potential links is required. The accuracy of the proposed estimator appears to be bounded by the accuracy of the dual system estimator with perfect linkage and it diminishes as the discriminatory power of the linkage variables decreases.