The effects of handling outliers on the performance of bankruptcy prediction models

Ratio type financial indicators are the most popular explanatory variables in bankruptcy prediction models. These measures often exhibit heavily skewed distribution because of the presence of outliers. In the absence of clear definition of outliers, ad hoc approaches can be found in the literature for identifying and handling extreme values. However, it is not clear how these different approaches can affect the predictive power of models. There seems to be consensus in the literature on the necessity of handling outliers, at the same time, it is not clear how to define extreme values to be handled in order to maximize the predictive power of models. There are two possible ways to reduce the bias originating from outliers: omission and winsorization. Since the first approach has been examined previously in the literature, we turn our attention to the latter. We applied the most popular classification methodologies in this field: discriminant analysis, logistic regression, decision trees (CHAID and CART) and neural networks (multilayer perceptron). We assessed the predictive power of models in the framework of tenfold stratified crossvalidation and area under the ROC curve. We analyzed the effect of winsorization at 1, 3 and 5% and at 2 and 3 standard deviations, furthermore we discretized the range of each variable by the CHAID method and used the ordinal measures so obtained instead of the original financial ratios. We found that this latter data preprocessing approach is the most effective in the case of our dataset. In order to check the robustness of our results, we carried out the same empirical research on the publicly available Polish bankruptcy dataset from the UCI Machine Learning Repository. We obtained very similar results on both datasets, which indicates that the CHAID-based categorization of financial ratios is an effective way of handling outliers with respect to the predictive performance of bankruptcy prediction models.     

Identifying significant characteristics of organic milk consumers: a CART analysis of an artefactual field experiment

The organic dairy category is one of the fastest growing categories of organic foods in the US. Organic milk consumers generally cite perceived health benefits and lower risk of food contamination, as well as perceived superior quality and environmental sustainability of organic farming methods, as the major motivations for preference of organic over conventional milk. While the attributes of organic milk that are valued by consumers are fairly well-known, more ambiguity exists regarding the demographic characteristics of the typical organic milk consumer. This research makes use of experimental data from 148 adult participants and use a Classification and Regression Tree (CART) analysis, a nonparametric modelling approach, to identify how Willingness-to-Pay (WTP) for organic milk varies with the demographic profile of experiment participants. The study finds that perceived taste of organic milk, concern for the risk of consuming conventional milk, being a primary shopper, and the quantity of milk consumed are the major factors that separate experiment participants into groups with high and low WTP for organic milk.     

Segmenting Cruise Passengers Visiting Uruguay: a Factor–Cluster Analysis

The study aims to provide a better understanding of cruise travel from passengers' characteristics and experience in two ports of call in Uruguay. A multivariate market segmentation analysis was used, on the basis of 5151 survey data collected during the 2008–2009 and 2009–2010 seasons. A correspondence analysis revealed the underlying latent factors in the set of variables. A hierarchical clustering from correspondence analysis segmented the sample into homogeneous groups. Finally, a decision tree highlighted the most predictive variables for each cluster. The study identifies distinct segments by country of residence, occupation, locations visited in Uruguay, satisfaction and previous visits to the country. Copyright © 2012 John Wiley & Sons, Ltd.     

基于多时相HJ卫星的渭干河-库车河绿洲主要农作物种植信息提取

在3S技术的支持下全面、快速、客观地监测农作物种植信息,对于正确的把握该区域的农业结构和布局,进行作物种植空间格局的调整和优化有着十分重要的意义.本文以渭干河-库车河洲绿洲为研究区,根据农作物的物候规律和季相节律的差异性特点,选取2012年的3景不同时相的HJ卫星CCD遥感数据,ENVI下基于CART算法的决策树规则自动提取主要农作物覆盖信息,然后以野外GPS调查点为依据,对决策树方法预分类结果进行修正,成功提取了研究区的玉米、棉花和小麦的种植面积,总体精度达到了91.73％.结果表明HJ卫星CCD影像可以很好地应用于农作物提取,而且CART算法的分类精度较高,能较好地反映作物的分布状况,可为该地区主要作物种植结构调整提供一定的依据.     

Default Predictors in Retail Credit Scoring: Evidence from Czech Banking Data

Credit to the private sector has risen rapidly in European emerging markets, but its risk evaluation has been largely neglected. Using retail-loan banking data from the Czech Republic, we construct two credit risk models based on logistic regression and classification and regression trees. Both methods are comparably efficient and detect similar financial and socioeconomic variables as the key determinants of default behavior. We also construct a model without the most important financial variable (amount of resources), which performs very well. This way, we confirm significance of sociodemographic variables and link our results with specific issues characteristic to new EU members.     

中国地方政府支出与居民消费关系的区域差异研究

本文利用分类回归树(CART)方法,根据中国地方政府支出与居民消费关系的差异将中国31个省(市、区)分为两组,分别对两组区域进行面板数据分析。结果表明:政府支出增长对居民消费的影响在第一组区域具有显著的引致效应,而在第二组区域则表现为挤出效应。在第一组,农业支出、政策性补贴和政府消费的增长对居民消费具有较大的正向影响,但只能在当期发挥作用;在第二组,则是基本建设支出和科学技术支出的增长显著促进了居民消费,且显示出较好的持续性。应根据不同的区域特征采取针对性措施,以有效拉动居民消费增长。     

Drivers influencing farmer decisions for adopting organic or conventional coffee management practices

Colombia is one of the world’s most important producers of Arabica coffee (Coffea arabica), whose coffee-growing zone coincides with a biogeographic hotspot of biodiversity. Given that coffee agroecosystems are grown by both organic and conventional schemes of management in Santander, a region which produces coffees with specialist distinctive flavours, this study aims to better understand the factors that influence the adoption of these different schemes of management. A combination of ethnographic techniques and quantitative methods were used to examine the predominant drivers of adoption and revealed farmer perceptions associated with coffee farming, and the complexity of interacting factors, that surround their decision making. The results of qualitative analysis suggests that social identity of coffee growers, the existence of farming spaces (lived, perceived, rationalised), the influence of coffee institutions, attitudes about management practices, and social relations of production, all play an important role in the process of decision making. In quantitative terms, we identified 18 socioeconomic drivers, some with interacting effects that had significant influence on the decision to adopt either organic or conventional practices. In particular, at local scale, important factors were technology availability, the type of landowner, formal education of farmers, the role of institutions, membership of community organisations, farm size, coffee productivity and the number of coffee plots per farm. Likewise, economic drivers, such as crop profitability, determined how farmers are involved in trade and market networks at broad regional, national, and international spatial scales. By adopting a more integrated approach, combining qualitative and quantitative methodologies, we characterised the complexity of factors that influencing adoption of coffee management schemes and show that not only financial factors but also a variety of other social factors drive farmer decision making. Identifying the most influential behavioural drivers provides policy with opportunities to better support farmer livelihoods.     

Predicting and explaining absenteeism risk in hospital patients before and during COVID-19

In order to address one of the most challenging problems in hospital management – patients’ absenteeism without prior notice – this study analyses the risk factors associated with this event. To this end, through real data from a hospital located in the North of Portugal, a prediction model previously validated in the literature is used to infer absenteeism risk factors, and an explainable model is proposed, based on a modified CART algorithm. The latter intends to generate a human-interpretable explanation for patient absenteeism, and its implementation is described in detail. Furthermore, given the significant impact, the COVID-19 pandemic had on hospital management, a comparison between patients’ profiles upon absenteeism before and during the COVID-19 pandemic situation is performed. Results obtained differ between hospital specialities and time periods meaning that patient profiles on absenteeism change during pandemic periods and within specialities.