Finite Mixture Model Approach Research Articles

A Bayesian finite mixture model approach to evaluate dichotomization method for correlated ELISA tests

In diagnostic accuracy studies, a commonly employed approach involves dichotomizing continuous data and subsequently analyzing them using a Bayesian latent class model (BLCM), often relying on binomial or multinomial distributions, rather than preserving their continuous nature. However, this procedure can inadvertently lead to less reliable outcomes due to the inherent loss of information when converting the original continuous measurements into binary values. Through comprehensive simulations, we demonstrated the limitations and disadvantages of dichotomizing continuous biomarkers from two correlated tests. Our findings highlighted notable disparities between the true values and the model estimates as a result of dichotomization. We discovered the crucial significance of selecting a reference test with high diagnostic accuracy in test evaluation in order to obtain reliable estimates of test accuracy and prevalences. Our study served as a call to action for veterinary researchers to exercise caution when utilizing dichotomization.

Modeling red-light running behavior using high-resolution event-based data: a finite mixture modeling approach

To effectively reduce the number of red-light violations and crashes, it is crucial to explore RLR behavior at local intersections, understand the contributing factors, and identify the riskiest intersections by estimating RLR frequency. In this study, a finite mixture modeling method was utilized to understand the contributing factors to RLR behavior and estimate this violating behavior. To develop the RLR estimation models, performance metrics and signal phasing data were collected from the Automated Traffic Signal Performance Measures (ATSPMs) system in two jurisdictions in Arizona: Pima County and the Town of Marana. The results from calibrated models showed that an increase in traffic flow, intersection delay, number of approach lanes, and split failure is associated with an increase in the likelihood of observing red-light violations. In addition, it was found that an increase in cycle length is associated with a decrease in the likelihood of observing the red-light violation. The results of comparing the proposed RLR estimation method with several conventional methods, the Poisson Generalized Linear Model (PGLM), Zero-inflated Poisson Regression Model (ZIPM), and Zero-inflated Negative Binomial Regression Model (ZINB) showed the proposed method outperforms all the models in terms of both model fit and accuracy. The application of the proposed method could be used to analyze the intersections with the highest number of red-light violations. Furthermore, the presented transferability results can be advantageous to transportation agencies within Arizona and urban areas with similar characteristics by providing insight into which model specifications may provide the best RLR estimation accuracy.

Classification, segmentation and correlation of zoned minerals

Minerals exhibit compositional zoning patterns that can be related to changes in the environment in which they grew. Using statistical methods that have been designed to segment optical images, we have developed a procedure to segment zoned crystals using elemental maps. For a single mineral phase, compositional zones in individual crystals are correlated between multiple crystals. This allows us to quantify the complexity and variability of chemical zoning between different geological samples. Specifically, we employ a simple linear iterative clustering algorithm, which splits the chemical maps into spatially constrained regions of similar chemistry. The result is a texturally segmented crystal, akin to what would be identified by the human eye. To aid the segmentation and correlation of zones, we also introduce a new method to classify multiple mineral phases within a single thin section. This is based on a finite mixture model approach, which proves very effective in removing mixed pixels that will only introduce noise into the segmentation. We provide an example using the mineral phase plagioclase. Using two contemporaneous samples from an eruptive unit on the island of St. Kitts (Eastern Caribbean) we show that a volcanic bomb (~10 cm) and a scoria clast (~2 cm) have similar rim compositions but distinctly different core compositions. Our methodology will enable a statistical characterization of 2D complexity of crystals in a variety of different geo-scientific disciplines. This will allow the genesis of different mineral phases to be directly compared.

Website Defacer Classification: A Finite Mixture Model Approach

Hackers often engage in website defacement early in their criminal careers to establish a reputation. Some hackers become increasingly prolific and launch a large number of attacks against their targets, whereas others only launch a few attacks before eventually desisting from a life of crime. A better understanding of why some hackers launch a large number of attacks, while others do not, will assist in the implementation of targeted intervention strategies. Therefore, the current study, using a sample of 119 active hackers, seeks to answer two research questions: (1) Are there different groups of website defacers based on attack volume? (2) Which observed hacker-level characteristics can be used to predict latent class membership? We find that two unique groups of website defacers exist: low-volume defacers (69%) and high-volume defacers (31%). Social media presence, the content of the defacement, and the type of defacement are all predictive of latent class membership. Policy implications are discussed.

Application of Bayesian Hierarchical Negative Binomial Finite Mixture Model for Cost-Benefit Analysis of Barriers Optimization, Accounting for Severe Heterogeneity

The Wyoming Department of Transportation (WYDOT) initiated a project to optimize the heights of barriers that are not satisfying the barrier design criteria, while prioritizing them based on an ability to achieve higher monetary benefits. The equivalent property damage only (EPDO) was used in this study to account for both aspects of crash frequency and severity. Data of this type are known to have overdispersion, that is having a variance greater than the mean. Thus, a negative binomial model was implemented to address the over-dispersion issue of the dataset. Another challenge of the dataset used in this study was the heterogeneity of the dataset. The data heterogeneity resulted from various factors such as data being aggregated across two highway systems, and the presence of two barrier types in the whole state. Thus, it is not practical to assign a subjective hierarchy such as a highway system or barrier types to address the issue of severe heterogeneity in the dataset. Under these conditions, a finite mixture model (FMM) was implemented to find a best distribution parameter to characterize the observations. With this technique, after the optimum number of mixtures was identified, those clusters were assigned to various observations. However, previous studies mostly employed just the finite mixture model (FMM), with various distributions, to account for unobserved heterogeneity. The problem with the FMM approach is that it results in a loss of information: for instance, it would come up with N number of equations, where each result would use only part of the whole dataset. On the other hand, some studies used a subjective hierarchy to account for the heterogeneity in the dataset, such as the effect of seasonality or highway system; however, those subjective hierarchies might not account for the optimum heterogeneity in the dataset. Thus, we implement a new methodology, the Bayesian Hierarchical Finite Mixture (BHFMM) to employ the FMM without losing information, while also accounting for the heterogeneity in the dataset, by considering objective and unbiased hierarchies. As the Bayesian technique has the shortcoming of labeling the observations due to label switching; the FMM parameters were estimated by maximum likelihood technique. Results of the identified model were converted to an equation for implementation of machine learning techniques. The heights were optimized to an optimal value and the EPDO was predicted based on the changes. The results of the cost–benefit analysis indicated that after spending about 4 million dollars, the WYDOT would not only recover the expenses, but could also expect to save more than $4 million additional dollars through traffic barrier crash reduction.

Improving revenue performance with third-degree price discrimination in the cruise industry

Abstract The leisure cruise industry has enjoyed high levels of growth for nearly five decades due in part to traveler interest in the cruise experience, but also to relatively lower pricing. Although revenue management of cruise fares is now standard practice, there are untapped opportunities to improve yields through data-driven market segmentation and third-degree price discrimination. This paper uses a finite mixture modeling approach to develop, empirically validate, and compare pricing models. By unveiling segments of travelers based on individual attributes, third-degree price discrimination can improve target marketing, the timing and appeal of price discounts, and the matching of variable demand with fixed, though differentiated, room supply. Empirical results from running thousands of simulations with pricing data from one of the world’s largest cruise lines show that the segmentation analysis using third-degree price discrimination can increase fare revenue more than four percent. The modeling approach used in this research extends the emerging literature on revenue management in the cruise industry and offers meaningful managerial implications for advanced pricing tactics and revenue management.

Regulatory focus and decision rules: Are prevention-focused consumers regret minimizers?

The Random Regret Minimization (RRM) discrete choice model, which assumes the emotional rationality of a decision maker, is introduced into consumer decision making. The potential consumer-specific drivers of the decision rules between utility maximization and regret minimization are scrutinized to better understand consumer decision-making processes. Drawing on Regulatory Focus Theory (RFT), we predict that chronically prevention-focused consumers are more likely to be regret minimizers, as they are more inclined to minimize negative outcomes, and that chronically promotion-focused consumers are more likely to be utility maximizers, as they are more inclined to maximize their opportunities and outcomes. Employing a finite mixture modeling approach, we find empirical support for the hypotheses. We also find a significant portion of regret-minimizing consumers showing different preferences for product design. The empirical findings highlight the importance of incorporating consumer segments with different decision rules as different marketing strategies may be required for product design and positioning.

Measuring neighbourhood social and economic change for urban health studies

Neighbourhood change is a complex phenomenon. To study its consequences for health outcomes, we developed a measure of neighbourhood social and economic change for all census tracts ( n = 2272) in the entire city of Madrid (Spain) in two epochs (2005–2009 and 2009–2013). We used a finite mixture modelling approach with 16 indicators from several administrative sources. We found four types of neighbourhoods: Decreasing Socioeconomic Status (SES) areas with increased diversity and decreased socioeconomic status; New Housing/Gentrification areas with high residential mobility, new housing construction and with markers of gentrification in the crisis epoch; Increasing SES areas with increased socioeconomic status and decreased diversity; and Aging areas with an aging population, low residential mobility and no new construction. We describe the baseline predictors of these types of change, finding that there is a potential widening of socioeconomic gaps, as Increasing SES areas start with higher SES, and Decreasing SES areas start with lower SES. We found a change in the spatial distribution of these types between the first and second epochs, as New Housing/Gentrification areas became more common in the centre of the city. We discuss two potential applications of this type of model to the study of the consequences of residential environment changes for health determinants and health outcomes, with a particular emphasis on retail food environments and diabetes incidence.

Does DRG funding encourage hospital specialization? Evidence from the Italian National Health Service.

The impact of diagnosis-related group (DRG)-based funding has been analyzed along a wide range of dimensions. Its effects on hospital specialization, however, have been investigated only sparsely. This paper examines such effects in the context of the Italian National Health Service, where decentralization has produced a significant degree of variation in funding arrangements. To this end, a 9-year panel data set covering 762 Italian public and private hospitals was analyzed using a finite mixture model approach. Hospital specialization was measured by the internal Herfindahl-Hirschman Index. Three variables were introduced as proxies for the choices made by Italian Regions with respect to the development and use of their DRG systems. The best finite mixture model identified three groups of hospitals, two of which sizeable. Of these, one included nearly all public hospitals, while the other was composed almost exclusively of small and medium-sized investor-owned hospitals. Averagely, private and smaller hospitals showed a stronger tendency to specialize over time. The positive impact of DRG funding on the hospitals' propensity to specialize found only limited empirical support. Moreover, it emerged as comparatively much smaller for public hospitals vis à vis private ones.

A finite mixture modeling approach to examine New York City bicycle sharing system (CitiBike) users’ destination preferences

Given the recent growth of bicycle-sharing systems (BSS) around the world, it is of interest to BSS operators/analysts to identify contributing factors that influence individuals’ decision processes in adoption and usage of bicycle-sharing systems. The current study contributes to research on BSS by examining user behavior at a trip level. Specifically, we study the decision process involved in identifying destination locations after picking up the bicycle at a BSS station. In the traditional destination/location choice approaches, the model frameworks implicitly assume that the influence of exogenous factors on the destination preferences is constant across the entire population. We propose a finite mixture multinomial logit (FMMNL) model that accommodates such heterogeneity by probabilistically assigning trips to different segments and estimate segment-specific destination choice models for each segment. Unlike the traditional destination choice based multinomial logit (MNL) model or mixed multinomial logit (MMNL), in an FMMNL model, we can consider the effect of fixed attributes across destinations such as users’ or origins’ attributes in the decision process. Using data from New York City bicycle-sharing system (CitiBike) for 2014, we develop separate models for members and non-members. We validate our models using hold-out samples and compare our proposed FMMNL model results with the traditional MNL and MMNL model results. The proposed FMMNL model provides better results in terms of goodness of fit measures, explanatory power and prediction performance.

Are all Homeowners Willing to Pay for Better Schools? ─ Evidence from a Finite Mixture Model Approach

School quality indicators such as student test scores have been shown to be capitalized into the value of local homes. The presence of households with different preferences for education implies that the implicit price of improvements in school quality might vary even within a region. In this paper, we employ a finite mixture model (FMM) to capture unobserved heterogeneity in household preferences. Using school quality and residential property sales data from Pitt County, North Carolina, we find evidence of two subpopulations of houses, where the prices for one group are virtually invariant to school quality. Consistent with recent research by Davis et al. (2017) these results indicate that heterogeneous valuation of educational quality by households with different socio-economic backgrounds should be taken into consideration when devising policies targeted at the local level.

Stochastic characterization of wind field characteristics of an arch bridge instrumented with structural health monitoring system

This paper aims to conduct a stochastic characterization of wind field characteristics nearby an arch bridge based on long-term monitoring data from an instrumented structural health monitoring (SHM) system. The fluctuating wind characteristics are first presented by analyzing the real-time wind monitoring data. A genetic algorithm (GA)-based finite mixture modeling approach is proposed to formulate the joint distribution of the wind speed and direction. For the probability density function (PDF) of the wind speed, a two-parameter Weibull distribution is applied, and a von Mises distribution is selected to present the PDF of the wind direction. The parameters of finite mixture models are estimated by the GA-based parameter estimation method. The effectiveness of the proposed direct probabilistic modeling approach is validated by use of one-year of wind monitoring data, and compared with the traditional angular-linear (AL) distribution-based indirect modeling approach in terms of the Akaike’s information criterion (AIC), Bayesian information criterion (BIC) and R2 value. Results indicate that the proposed GA-based finite mixture modeling approach fits the measured data better than the AL distribution-based indirect modeling approach. In addition, the joint distribution of the wind speed and direction will facilitate the wind-resistant design and wind-induced fatigue assessment of long-span bridges.

Neighborhood Typologies Associated with Alcohol Use among Adults in Their 30s: a Finite Mixture Modeling Approach.

There has been increasing interest in how neighborhood context may be associated with alcohol use. This study uses finite mixture modeling to empirically identify distinct neighborhood subtypes according to patterns of clustering of multiple neighborhood characteristics and examine whether these subtypes are associated with alcohol use. Neighborhoods were 303 census block groups in the greater Seattle, WA, area where 531 adults participating in an ongoing longitudinal study were residing in 2008. Neighborhood characteristics used to identify neighborhood subtypes included concentration of poverty, racial composition, neighborhood disorganization, and availability of on-premise alcohol outlets and off-premise hard liquor stores. Finite mixture models were used to identify latent neighborhood subtypes, and regression models with cluster robust standard errors examined associations between neighborhood subtypes and individual-level typical weekly drinking and number of past-year binge drinking episodes. Five neighborhood subtypes were identified. These subtypes could be primarily characterized as (1) high socioeconomic disadvantage, (2) moderate disadvantage, (3) low disadvantage, (4) low poverty and high disorganization, and (5) high alcohol availability. Adjusted for covariates, adults living in neighborhoods characterized by high disadvantage reported the highest levels of typical drinking and binge drinking compared to those from other neighborhood subtypes. Neighborhood subtypes derived from finite mixture models may represent meaningful categories that can help identify residential areas at elevated risk for alcohol misuse.

What Matters Most to People? Evidence from the OECD Better Life Index Users’ Responses

The OECD Better Life Index is an interactive composite index that aggregates a country’s well-being outcomes through the weights defined by online users. This paper analyses these weights by analysing the responses given by close to 88,000 users since 2011 to date. The contribution of this paper is threefold. First, it investigates the factors shaping users’ preferences over a set of 11 well-being dimensions, while most of the previous empirical works in the area have focused on factors affecting support for a specific well-being domain (e.g. redistribution, environmental concerns) at a time. Second, it provides insights into users’ preferences for a large group of countries, which differ in terms of culture and living conditions. Third, a finite mixture model (FMM) approach is used to test for heterogeneity in the effect of satisfaction levels on the weight attached to a given BLI dimension across sub-population groups. Various empirical models are used to identify responses’ patterns and see whether they can be accounted for respondents’ characteristics and their perceived level of well-being. The paper finds that health, education and life satisfaction are the aspects that matter the most in OECD countries. Descriptive statistics show that men assign more importance to material conditions than women; while women in general value quality of life more than men. Environment, housing, civic engagement, safety and health become more important with age, while life satisfaction, education, work-life balance, jobs and income are particularly important for those younger than 35. There are also regional patterns in users’ findings, for instance civic engagement is particularly important in South America, while safety and work-life balance matter tremendously in Asia-Pacific. Furthermore, an additional analysis carried out on a subset of observations finds that for several well-being dimensions (i.e. jobs, housing, community, health, education, civic engagement, safety, life satisfaction and work-life balance) there is a positive and linear relationship between individual preferences and self-reported satisfaction in those dimensions. Finally, the check for heterogeneity in the relationship of satisfaction to preferences in well-being dimensions, via an FMM analysis, reveals that, in the case of income and education, two classes of individuals with distinct effects of satisfaction levels on preferences are identified.

Finite mixture modeling approach for developing crash modification factors in highway safety analysis

This study aimed to investigate the relative performance of two models (negative binomial (NB) model and two-component finite mixture of negative binomial models (FMNB-2)) in terms of developing crash modification factors (CMFs). Crash data on rural multilane divided highways in California and Texas were modeled with the two models, and crash modification functions (CMFunctions) were derived. The resultant CMFunction estimated from the FMNB-2 model showed several good properties over that from the NB model. First, the safety effect of a covariate was better reflected by the CMFunction developed using the FMNB-2 model, since the model takes into account the differential responsiveness of crash frequency to the covariate. Second, the CMFunction derived from the FMNB-2 model is able to capture nonlinear relationships between covariate and safety. Finally, following the same concept as those for NB models, the combined CMFs of multiple treatments were estimated using the FMNB-2 model. The results indicated that they are not the simple multiplicative of single ones (i.e., their safety effects are not independent under FMNB-2 models). Adjustment Factors (AFs) were then developed. It is revealed that current Highway Safety Manual’s method could over- or under-estimate the combined CMFs under particular combination of covariates. Safety analysts are encouraged to consider using the FMNB-2 models for developing CMFs and AFs.

Abstract 4431: Tandem profiling of 5-methyl and 5-hydroxymethylcytosine in glioblastoma identifies tumor-specific distribution of 5-hydroxymethylcytosine and associations with patient survival

Abstract Glioblastomas exhibit widespread molecular alterations including a highly distorted epigenome. Genome-wide assessment of epigenetic alterations to cytosines in glioblastoma has largely been limited to 5-methylcytosine (5mC). 5-hydroxymethylcytosine (5hmC) is an additional cytosine modification that is deregulated in cancer, though the extent and context of these alterations at single cytosine resolution remain unclear. Through parallel processing of DNA with bisulfite and oxidative bisulfite treatments on Illumina 450K arrays we resolved both 5mC and 5hmC in 30 glioblastomas. We developed and applied a novel technique for estimating 5mC, 5hmC, and unmethylated proportions from array data to glioblastoma tissues and compare with normal brain tissue. Genomic distribution of 5hmC was associated with features of transcription despite the glioblastoma genome being relatively depleted of 5hmC. When integrating 5mC and 5hmC data using a Gaussian finite mixture model approach, we observed significant associations between 5hmC levels and gene-sets involved in immune and RNA regulatory processes while high 5mC classes were associated with receptor-mediated processes and cancer gene-sets. Significant differences in patient survival were observed among classes of 5hmC obtained from a recursively partitioned mixture model. Gene expression across epigenetic enzymes including methyltransferases and TETs was also investigated for relation with patterns and levels of 5hmC and 5mC. Analyses of elevated 5hmC with alternative splicing events that may promote disease progression are underway. Together, our results provide a genome-wide map of 5hmC in glioblastoma, indicate that transcription patterns as well as disrupted processes are associated with 5hmC patterns, and highlight the potential clinical prognostic utility of 5hmC in cancer. Citation Format: Kevin C. Johnson, E. Andres Houseman, Jessica E. King, Camilo E. Fadul, Brock C. Christensen. Tandem profiling of 5-methyl and 5-hydroxymethylcytosine in glioblastoma identifies tumor-specific distribution of 5-hydroxymethylcytosine and associations with patient survival. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 4431.

Default and naive diversification heuristics in annuity choice

Retirement income stream products are difficult for consumers to choose because of their high perceived risk, irreversibility, high expenditure, little opportunity for social learning and distant consequences. Prior literature is unclear about consumers’ use of heuristics in decumulation decisions or whether sociodemographics can help identify vulnerable consumers. In the context of Australia’s retirement income arrangements, we examine choices of life annuities and phased withdrawal products, and identify use of default options and the diversification (1/ n or 50:50) heuristic using a novel finite mixture modelling approach. The innovative feature of this approach is that it captures the very specific allocation pattern associated with choices based on deterministic decision rules, namely pronounced spikes at the locations of the particular heuristics with little mass in their surroundings. We show that more than 30% of decumulation choices rely on these two heuristics, and that cognitive and product knowledge limitations contribute to using such heuristics. The results have implications for public policy on decumulation of retirement savings, regulation of product disclosures and providers of annuity and phased withdrawal products. More generally, our model has the potential to provide better understanding of the use of heuristics in consumer decisions.

Unobserved tax avoidance and the tax elasticity of FDI

This paper investigates the tax responsiveness of multinational firms’ investment decisions in foreign countries, distinguishing firms that are able to avoid taxes (avoiders) from those that are not (non-avoiders). From a theoretical point of view, the tax responsiveness of firms crucially depends on this distinction. Empirically, however, a firm's ability to avoid profit taxes is inherently unobservable to the researcher. To address this problem, we use a finite mixture modeling approach which allows us to distinguish avoiders from non-avoiders stochastically from a mixture of distributions of the two types of firms. Using panel data on the universe of foreign affiliates of German multinational firms over the years 1999–2010, we find that investments of tax avoiders do not respond to host-country profit taxes at all, while those of non-avoiders do. About 11% of the affiliates are estimated to be able to avoid taxes. These investments account for about 58% of the stock of foreign fixed assets held by German multinational firms abroad. A one-percentage-point increase in the statutory corporate profit tax rate of a host country is found to reduce the fixed assets of non-avoiders in that host country by 0.81%.

Default and 1/N Heuristics in Annuity Choice

Retirement income stream products are difficult for consumers to choose because of their high perceived risk, irreversibility, and high expenditure; little opportunity for social learning; and distant consequences. Yet many countries have shifted more responsibility for these choices to consumers, increasing their likelihood of using suboptimal heuristics without carefully evaluating alternatives. Prior literature is unclear about consumers’ use of heuristics in decu-mulation decisions or whether sociodemographics can help identify vulnerable consumers. The purpose of this paper is to address some of these gaps by examining choices of life annui-ties and phased withdrawal products and identifying use of default options and the diversifica-tion (1/n or 50:50) heuristic using a novel finite mixture modeling approach. We show that more than 30% of decumulation choices rely on these two heuristics and that cognitive and product knowledge limitations contribute to using such heuristics. The results have implica-tions for public policy on decumulation of retirement savings, regulation of product disclo-sures, and marketing of annuity and phased withdrawal products. More generally, the model we develop and apply has the potential to provide better understanding of the use of heuristics in consumer decisions.

Bayesian approach for mixture models with grouped data

Finite mixture modeling approach is widely used for the analysis of bimodal or multimodal data that are individually observed in many situations. However, in some applications, the analysis becomes substantially challenging as the available data are grouped into categories. In this work, we assume that the observed data are grouped into distinct non-overlapping intervals and follow a finite mixture of normal distributions. For the inference of the model parameters, we propose a parametric approach that accounts for the categorical features of the data. The main idea of our method is to impute the missing information of the original data through the Bayesian framework using the Gibbs sampling techniques. The proposed method was compared with the maximum likelihood approach, which uses the Expectation-Maximization algorithm for the estimation of the model parameters. It was also illustrated with an application to the Old Faithful geyser data.