Automatic Interaction Detection Algorithm Research Articles

Choices of medical institutions and associated factors in older patients with multimorbidity in stabilization period in China: A study based on logistic regression and decision tree model

AbstractBackgroundAs China's population ages, its disease spectrum is changing, and the coexistence of multiple chronic diseases has become the norm with respect to the health status of its elderly population. However, the health institution choices of older patients with multimorbidity in stabilization period remains underresearched. This study investigate the factors influencing the choices of older patients with multimorbidity to provide references for the rational allocation of healthcare resources.MethodsA multistage, stratified, whole‐group random‐sampling method was used to select eligible older patients from September to December of 2022 who attended the Community Health Service Center of Guangdong Province. We adopted a self‐designed questionnaire to collect patients' general, disease‐related, social‐support information, their intention to choose a healthcare provider. A binary logistic regression and decision tree model based on the Chi‐squared automatic interaction detector algorithm were implemented to analyze the associated factors involved.ResultsA total of 998 patients in stabilization period were included in the study, of which 593 (59.42%) chose hospital and 405 (40.58%) chose primary care. Our binary logistic regression results revealed that age, sex, individual average annual income, educational level, self‐reported health status, activities of daily living, alcohol consumption, family doctor contracting, and family supervision of medication or exercise were the principal factors influencing the choice of medical institutions for older patients with multimorbidity (p < 0.05). The decision‐tree model reflected three levels and 11 nodes, and we screened a total of four influencing factors: activities of daily living, age, a family doctor contract, and patient sex. The data showed that the logistic regression model possessed an accuracy of 72.9% and that the decision tree model exhibited an accuracy of 68.7%. Prediction using the binary logistic regression was thus statistically superior to the categorical decision‐tree model based on the Chi‐squared automatic interaction detector algorithm (Z = 3.238, p = 0.001).ConclusionMore than half of older patients with multimorbidity in stabilization period chose hospitals for healthcare. Efforts should be made to improve the quality of healthcare services and increase the medical contracting rate and recognition of family doctors so as to attract older patients with multimorbidity to primary medical institutions.

Who wants to know? Older adults’ interest in housing information

ABSTRACTNearly 80% of adults aged 45+ reported wanting to remain in their homes and communities as they age and stay at home even when they experience physical and cognitive declines. As the population of adults age 65+ increases, the market of mature consumers will also increase placing expectations and demands on industries including healthcare, housing, and service. Decision makers in these increasingly competitive arenas will need to position their missions and goals to align with these changing needs. The goal of this study is to identify characteristics of adults aged 45+ who are interested in community-based housing options including home repair/modification programs, reverse mortgage options, retirement communities, and assisted living facilities. Data were collected through a nationwide telephone survey to identify the current and future information needs of adults aged 45+. Andersen’s behavioral model was used as a framework for selecting variables to describe housing information seekers. Using classification and regression tree analysis with the chi-squared automatic interaction detector algorithm to overcome issues of multicollinearity within the data, findings indicated that perception of health interference in daily living, age, gender, and living arrangement was the significant indicators for interest in obtaining housing information.

Integration of Monte Carlo simulation and chi-square automatic interaction detector algorithm for modelling and estimation of conventional power plant construction volumes

This study presents an integrated approach based on chi-square automatic interaction detector (CHAID) algorithm, computer simulation and statistical tools for optimum modelling and estimation of power plant construction volumes. In this context, modelling applications in estimation of structural volumes in power plants is described. The required information for cabling volumes such as low-pressure and cable platter are collected from an actual power plant construction company. Since some of these projects cannot respond to essential data volume for modelling, Monte Carlo simulation method is used in this study to generate the required data. Moreover, based on existing information and experts’ judgement, sufficient data will be generated. Then, using the simulated data, CHAID is introduced and used for estimation of construction volumes. This is the first study that integrates simulation and CHAID for optimum modelling of conventional power plant construction volumes.

Singapore credit cardholders: ownership, usage patterns, and perceptions

PurposeThe purpose of this study is to analyze Singapore's diverse cardholders in search of variations among demographic groups, credit card profiles, and their perceptions with regard to credit card ownership and use. It then aims to discuss possible reasons governing Singaporeans' credit card ownership and use.Design/methodology/approachA survey was conducted, decision trees were then constructed using Chi‐square automatic interaction detection algorithm (CHAID) and SPSS software AnswerTree to examine the association between the number of credit cards (target variable) and the demographic characteristics, perceptions and other credit card‐related variables.FindingsThe number of credit cards was found to be significantly influenced by income and gender as well as perceptions that include “credit card leads to overspending”, “savings as payment source”, “unreasonable interest rates”, “credit card as status symbol”. The number of credit cards was also affected by credit card‐related variables such as missing payments sometimes, frequency of use, entertainment expenditures, and petrol purchase.Practical implicationsThis research provides an in‐depth understanding of Singaporean multiple cardholders, thus it is useful in designing marketing strategies for card‐issuers as well as anti‐debt strategies for policy‐makers in Singapore.Originality/valueDespite the importance of consumer credit, virtually no literature or research exists on the ownership and use of credit cards in Singapore, so this paper intends to close this gap. Further, by combining the demographics, cardholders' profiles and usage patterns with the respondents' perceptions concerning credit card ownership and use, our study offers a richer analysis to explain consumer behavior than previous literatures.