随机森林方法及其在客户流失预测中的应用研究

在全球化的市场竞争中,企业如何利用现有资源,提高客户满意度,保住现有客户,已成为企业面临的主要问题,客户流失预测越来越受到企业关注。本文针对实际客户流失数据中正负样本数量不平衡而且数据量大的特点,提出一种改进的平衡随机森林算法,并将其应用于某商业银行的客户流失预测。实际数据集测试结果表明,与传统的预测算法比较,这种算法集成了抽样技术和代价敏感学习的优点,适合解决大数据集和不平衡数据,具有更高的精确度。     

基于随机便利收益的不完全市场商品期货定价研究

商品的弱流动性导致了商品期货市场的不完全性。本文在现有商品便利收益期货定价模型的基础上,考虑了商品期货市场的不完全性及现货价格的Poisson跳跃过程,运用随机贴现因子与随机便利收益将商品现货价格与期货价格连接,提出了随机便利收益下期货市场不完全性的期货定价模型。为检验模型的适用性,利用上海期货交易所五只铜期货合约的交易数据对模型进行实证,并估计不完全参数,结果表明由于不完全性而导致的期货市场部分波动应主要归因于商品的随机便利收益。     

Geospatial Modeling of Containment Probability for Escaped Wildfires in a Mediterranean Region

Despite escalating expenditures in firefighting, extreme fire events continue to pose a major threat to ecosystem services and human communities in Mediterranean areas. Developing a safe and effective fire response is paramount to efficiently restrict fire spread, reduce negative effects to natural values, prevent residential housing losses, and avoid causalties. Though current fire policies in most countries demand full suppression, few studies have attempted to identify the strategic locations where firefighting efforts would likely contain catastrophic fire events. The success in containing those fires that escape initial attack is determined by diverse structural factors such as ground accessibility, airborne support, barriers to surface fire spread, and vegetation impedance. In this study, we predicted the success in fire containment across Catalonia (northeastern Spain) using a model generated with random forest from detailed geospatial data and a set of 73 fire perimeters for the period 2008–2016. The model attained a high predictive performance (AUC = 0.88), and the results were provided at fine resolution (25 m) for the entire study area (32,108 km²). The highest success rates were found in agricultural plains along the nonburnable barriers such as major road corridors and largest rivers. Low levels of containment likelihood were predicted for dense forest lands and steep-relief mountainous areas. The results can assist in suppression resource pre-positioning and extended attack decision making, but also in strategic fuels management oriented at creating defensive locations and fragmenting the landscape in operational firefighting areas. Our modeling workflow and methods may serve as a baseline to generate locally adapted models in fire-prone areas elsewhere.     

Modelling ordinal assessments: fit is not sufficient

Assessments in ordered categories are ubiquitous in the social sciences. These assessments are assigned ordinal counts and analyzed with probabilistic models. If the counts fit the model, it is assumed that no unaccounted for factors govern the distribution and that it is a random error distribution. However, because tests of fit utilize parameter estimates from the data, the data may fit the model even when the modeled distributions cannot be random error distributions. This paper applies the additional criterion of strict unimodality, common to all random error distributions, to decide if a modeled distribution is not a random error distribution. However, not only are common random error distributions strictly unimodal, they are also strictly log-concave, a stronger form of unimodality which ensures smooth transitions between probabilities of adjacent counts. The paper shows that the operation for determining the strict unimodality also ensures that the distribution is locally strictly log-concave around the measure of the entity of assessment.     

Central limit theorems for functionals of large sample covariance matrix and mean vector in matrix‐variate location mixture of normal distributions

In this paper, we consider the asymptotic distributions of functionals of the sample covariance matrix and the sample mean vector obtained under the assumption that the matrix of observations has a matrix‐variate location mixture of normal distributions. The central limit theorem is derived for the product of the sample covariance matrix and the sample mean vector. Moreover, we consider the product of the inverse sample covariance matrix and the mean vector for which the central limit theorem is established as well. All results are obtained under the large‐dimensional asymptotic regime, where the dimension p and the sample size n approach infinity such that p/n→c ∈ [0, + ∞) when the sample covariance matrix does not need to be invertible and p/n→c ∈ [0,1) otherwise.     

Demographic and HIV status diversities as mechanisms of social integration and segregation among Black sexual and gender minorities enrolled in a community-based social network intervention

Drawing on a social integration and intersectionality framework, this study advances a dynamic network understanding of the mechanisms that enable differential patterns of within-group social integration and segregation among Black sexual and gender minorities (BSGM). Specifically, in a cohort of BSGM (18–35 years of age, n = 340) participating in a community-based network intervention for HIV prevention, we examine how sexual, gender, age, and HIV status diversities contribute to friendship formation and maintenance patterns over the 12-month study enrollment period. We found attenuated social integration (or social activity) among non-gay-identified and older BSGM and evidence of social segregation (or homophily) on the basis of sexual identity and age similarities. Accounting for the moderating effects of the intervention revealed that the attenuated integration of non-gay-identified and older BSGM were stronger for participants who received the peer leadership training, and integration challenges were also found for transgender BSGM who received the peer leadership training. Meanwhile, BSGM living with HIV who received the peer leadership training were significantly more integrated than their counterparts in the control arm. These findings help us understand the complicated social fabric among BSGM and the dynamics that interventions for this community may have to contend with or alter.     

C55. Quick and accurate rule of thumb for determining the critical value of the coefficient of correlation

Life table analysis techniques in epidemiology depend upon the asymptotic properties of the statistical test methods employed. In some instances, the statistical procedures indicate highly significant results which are, in reality, unjustified. The phenomenon may occur when the asymptotic methods are applied in situations where the cases of interest are few in number. This situation is illustrated by the 20 multiple myeloma deaths observed in the RERF Life Span Study cohort. A permutation test is applied to the life table data, although the test requires the false assumption that the censoring distribution is independent of the radiation dose. A simulation test is developed which does not require equal censoring, which has the same asymptotics as the usual test methods, and which is less likely to overestimate significance in small samples. It is found that both of these small-sample tests provide reasonable numerical solutions. In addition, the simulation test is recommended in general for analyzing life table data with unequal censoring. Finally, by using the small-sample tests, the frequency of death from multiple myeloma is shown to be positively associated with radiation dose (P<0.01).     

Relative efficienctes of sampling plans for selecting a small number of units from a rectangular region

We examine the efficiency of several sampling plans for use in certain agricultural, ecological and environmental studies. One concern for such studies is that plots that arephysically close might be more similar than distant plots. We considered sampling plansthat are designed to generate samples that represent the entire population while avoidingthe selection of units that provide essentially redundant information. All plans havethe property that they avoid the simultaneous selection of units that are, in some sense,neighboring units. By means of a simulation study, the efficiency of these plans iscompared to simple random Aampling Factors that influence the relative efficiencies areexamined. This is done for a number of different populations, representing variouspossible patterns for a response variable.     

The strong consistency of M-estimates in linear models with extended negatively dependent errors

In this paper, we first establish the strong convergence for weighted sums of extended negatively dependent (END) random variables. Based on the strong convergence and Bernstein inequality, we obtain the strong consistency of M-estimates of the regression parameters in a linear model for END random errors under some mild moment conditions. The results generalize and improve the ones obtained in the literature to the case of END random errors.