Chi-squared Automatic Interaction Detection Model Research Articles

Predicting mortality amongst Jordanian men with heart attacks using the chi-square automatic interaction detection model.

Background: One of the most complicated cardiovascular diseases in the world is heart attack. Since men are the most likely to develop cardiac diseases, accurate prediction of these conditions can help save lives in this population. This study proposed the Chi-Squared Automated Interactive Detection (CHAID) model as a prediction algorithm to forecast death versus life among men who might experience heart attacks. Methods: Data were extracted from the electronic health solution system in Jordan using a retrospective, predictive study. Between 2015 and 2021, information on men admitted to public hospitals in Jordan was gathered. Results: The CHAID algorithm had a higher accuracy of 93.72% and an area under the curve of 0.792, making it the best top model created to predict mortality among Jordanian men. It was discovered that among Jordanian men, governorates, age, pulse oximetry, medical diagnosis, pulse pressure, heart rate, systolic blood pressure, and pulse pressure were the most significant predicted risk factors of mortality from heart attack. Conclusion: With heart attack complaints as the primary risk factors that were predicted using machine learning algorithms like the CHAID model, demographic characteristics and hemodynamic readings were presented.

Using decision tree to predict non-suicidal self-injury among young adults: the role of depression, childhood maltreatment and recent bullying victimization.

Importance: Non-suicidal self-injury (NSSI) is a significant mental health issue requiring a deeper understanding of its underlying causes, such as childhood maltreatment, adult bullying victimization, and depression. Previous studies have not adequately addressed the cumulative risks of these factors on NSSI among college students. This population-based study investigates these cumulative risk factors.Design, setting, and participants: The cross-sectional study included 63 university's college students with a mean age of 19.6 years (N = 95,833).Main outcomes and measures: Two Chi-Square Automatic Interaction Detection (CHAID) decision tree models were used to classify subgroups based on childhood maltreatment and adult bullying victimization experiences and to investigate their cumulative risks of NSSI. Recursive partitioning algorithms determined each predictor variable's relative importance.Results: The CHAID model accurately predicted NSSI behaviours with an overall accuracy rate of 77.8% for individuals with clinically relevant depressive symptoms and 97.2% for those without. Among depressed individuals, childhood emotional abuse was the strongest NSSI predictor (Chi-Square, 650.747; adjusted P < .001), followed by sexual and physical abuse. For non-depressed individuals, emotional abuse in childhood was the strongest NSSI predictor (Chi-Square, 2084.171; adjusted P < .001), with sexual and verbal bullying in the past year representing the most significant proximal risks.Conclusions and relevance: Emotional abuse during childhood profoundly impacts individuals, increasing the risk of NSSI in both depressed and non-depressed individuals. Clinically relevant depressive symptoms have a moderating effect on the relationship between childhood maltreatment, adult bullying victimization, and NSSI. Identifying these factors can inform targeted interventions to prevent NSSI development among young adults.

Unaccompanied unstable housing among racially, ethnically, sexually, and gender diverse youth: Intersecting identities bearing the greatest burden.

Disparities in youth homelessness by racial/ethnic, sexual, and gender identities are well documented, though this literature lacks specificity regarding intersectional social identities of youth who are most likely to experience homelessness. Population-based cross-sectional data on youth from the 2019 Minnesota Student Survey (N = 80,456) were used to examine the relationship between parent caring and intersections of minoritized identities that experience the highest prevalence of two distinct types of unaccompanied unstable housing with expanded categories of sexual and gender identities. Exhaustive chi-square automatic interaction detection models revealed that low parent caring was the most common predictor of unaccompanied homelessness and running away, but there was important variation among youth of color at the intersection of sexual and gender identities. The findings reveal a more complex story of disparities in unaccompanied unstable housing among youth with multiple marginalized social identities and highlight the need to create culturally informed prevention and intervention strategies for parents of LGBTQ+ (lesbian, gay, bisexual, transgender, queer, and questioning) youth of color. The implications for prevention and intervention among subgroups with the highest prevalence are discussed in the context of interlocking systems of power and oppression. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Interaction of admission platelet count with current medications and the risk for chronic subdural recurrence.

Chronic subdural hematoma (cSDH) has a reported 10%-24% rate of recurrence after surgery, and prognostic models for recurrence have produced equivocal results. The objective of this study was to leverage a data mining algorithm, chi-square automatic interaction detection (CHAID), which can incorporate continuous, nominal, and binary data into a decision tree, to identify the most robust predictors of repeat surgery for cSDH patients. This was a retrospective cohort study of all patients with SDH from two level 1 trauma centers at a single institution. All patients underwent cSDH evacuation performed by 15 neurosurgeons between 2011 and 2020. The primary outcome was the rate of repeat surgery for recurrent cSDH following the initial evacuation. The authors used CHAID to identify relevant predictors of repeat surgery, including age, sex, comorbidities, postsurgical complications, platelet count prior to the first procedure, midline shift prior to the first procedure, hematoma volume, and preoperative use of anticoagulants, antiplatelets, or statins. Sixty (13.8%) of 435 study-eligible patients (average age 74.0 years) had a cSDH recurrence. These patients had 2.0 times greater odds of having used anticoagulants. The final CHAID model had an overall accuracy of 87.4% and an area under the curve of 0.76. According to the model, the predictor with the strongest association with cSDH recurrence was admission platelet count. Approximately 26% of patients (n = 23/87) with an admission platelet count < 157 × 109/L had a cSDH recurrence, whereas none of the 44 patients with admission platelets > 313 × 109/L had a recurrence. Approximately 17% of patients in the 157-313 × 109/L platelet group who had used preoperative statins required a second procedure, which was associated with a 2.3 times increased risk for repeat surgery compared to those who had not used statins preoperatively. Among those who had not used preoperative statins, a platelet count ≤ 179 × 109/L on admission for the first procedure was the strongest differentiator for a second surgery (n = 5/22 [23%]), which increased the risk of recurrence by 4.5 times. Among the patients using preoperative statins, the use of anticoagulants was the strongest differentiator for requiring repeat surgery (n = 11/33 [33%]). The described model identified platelet count on admission as the most important predictor of repeat cSDH surgery, followed by preoperative statin use and anticoagulant use. Critical cutoffs for platelet count were identified, which future studies should evaluate to determine if they are modifiable or reflective of underlying disease states.

Effective Prediction of Mortality by Heart Disease Among Women in Jordan Using the Chi-Squared Automatic Interaction Detection Model: Retrospective Validation Study.

Many current studies have claimed that the actual risk of heart disease among women is equal to that in men. Using a large machine learning algorithm (MLA) data set to predict mortality in women, data mining techniques have been used to identify significant aspects of variables that help in identifying the primary causes of mortality within this target category of the population. This study aims to predict mortality caused by heart disease among women, using an artificial intelligence technique-based MLA. A retrospective design was used to retrieve big data from the electronic health records of 2028 women with heart disease. Data were collected for Jordanian women who were admitted to public health hospitals from 2015 to the end of 2021. We checked the extracted data for noise, consistency issues, and missing values. After categorizing, organizing, and cleaning the extracted data, the redundant data were eliminated. Out of 9 artificial intelligence models, the Chi-squared Automatic Interaction Detection model had the highest accuracy (93.25%) and area under the curve (0.825) among the build models. The participants were 62.6 (SD 15.4) years old on average. Angina pectoris was the most frequent diagnosis in the women's extracted files (n=1,264,000, 62.3%), followed by congestive heart failure (n=764,000, 37.7%). Age, systolic blood pressure readings with a cutoff value of >187 mm Hg, medical diagnosis (women diagnosed with congestive heart failure were at a higher risk of death [n=31, 16.58%]), pulse pressure with a cutoff value of 98 mm Hg, and oxygen saturation (measured using pulse oximetry) with a cutoff value of 93% were the main predictors for death among women. To predict the outcomes in this study, we used big data that were extracted from the clinical variables from the electronic health records. The Chi-squared Automatic Interaction Detection model-an MLA-confirmed the precise identification of the key predictors of cardiovascular mortality among women and can be used as a practical tool for clinical prediction.

The relationship between depressive and anxious symptoms and school attendance among adolescents seeking psychological services in a public general hospital in China: a cross-sectional study

BackgroundSchool attendance problems (SAPs), whether absenteeism or dropout, are strongly associated with poor outcomes for adolescents. We examined multiple variables that influence SAPs to identify potential leverage points for improving school attendance.MethodsSelf-reported SAPs and demographic information was collected from 392 adolescents in adolescents presenting to the general hospital for psychological services. PHQ-9 and GAD-7 were applied to assess the severity of depressive and anxious symptoms. We constructed logistic regression analysis and the Chi-Square Automatic Interaction Detection (CHAID) segmentation analysis via SPSS Decision Tree to identifying risk factors for the development of SAPs in adolescents.ResultsSAPs were self-reported by 252 (64.3%) adolescents. The SAPs group and non-SAPs group showed a significant difference in age, PHQ9 total scores, GAD7 total scores, schools, siblings, residence, parental marital quality, general health, regular exercise, and regular diet. A post hoc comparison between the two groups showed that the frequency of SAPs was significantly higher in the moderately-severe and severe depressive groups compared with other three groups (none, mild, moderate). The frequency of SAPs in severe anxious groups was significantly different from the none-anxious group. According to the binary logistic regression analysis, the depressive severity, siblings, residence, marital quality of parents, general health, and regular diet were correlated with the SAPs among adolescents. The adjusted OR of SAPs according to moderately-severe depressive symptoms was 10.84 (95%CI: 1.967–59.742) and severe depressive symptoms was 6.659 (95%CI: 1.147–38.666). In the decision tree model, PHQ-9 severity was extracted as the first splitting variable, with regular exercise and residence as the second, and siblings as the third. The ROC curves for predicting SAPs showed a fair diagnostic accuracy of the model with AUCs of CHAID model (0.705,95%CI:0.652–0.759, P = 0.000) and logistic regression model (0.777,95%CI:0.729–0.824, P = 0.000).ConclusionOur study provides insights into the associations between depressive symptoms and poor school attendance and identifies a number of risk factors associated with SAPs. Effective intervention by mental health practitioners, more attention by policy makers, and further research in this area are urgently needed for adolescents.

Predictors of self-reported sexually transmitted infections among Turkish university students.

Sexually transmitted infections (STIs) are among the most crucial health problems that sexually active university students may encounter. This study aims to identify the predictors of self-reported STIs among university students. Study participants were 2,241 individuals who reported having had sexual intercourse among a cohort of 9,693 students surveyed from 21 Turkish universities. The age of participants ranged from 17 years to 28 years. The Chi-Square Automatic Interaction Detection (CHAID) analysis indicated that gender was the main predictor of self-reported STI. Also, number of partners and substance use were found as predictor variables for males. Finally, the classification accuracy of the CHAID model was found to be 95.3% within the sample. The present findings shed light on risk factors for STI acquisition, suggesting possible directions for tailoring future preventive interventions.

Assessment of Factors Contributing to Consumer Attitudes and Preferences toward Chantenay Carrots using the CHAID Model

Increasing the consumption of fresh vegetables to the recommended dietary guidelines by the World Health Organization (WHO) is the top priority for many national policy makers and horticulture businesses. One growth strategy which has been shown to promote vegetable intake is introducing different vegetable varieties into the market. This research analyses factors affecting consumer attitudes and willingness to purchase a new variety of carrots imported to Australia from the UK—the Chantenay carrot. To better inform marketing tactics for this new variety and to provide insights for predictions for uptake, a general carrot usage and attitude survey was conducted in four Australian states (N = 300). Results indicated that: sustainable packaging options; the size of carrots; cooking tips; taste; and place of origin were attitudinal triggering factors for purchasing fresh carrots in this sample. These factors were compared to the characteristics of the Chantenay variety using a CHi-squared Automatic Interaction Detection (CHAID) model to gauge the likelihood of acceptance and uptake of the variety within the larger Australian population. This study benefits marketing practitioners and horticultural businesses to develop a new vegetable variety that meets consumer preferences successfully.

A decision tree analysis to predict clinical outcome of minimally invasive lumbar decompression surgery for lumbar spinal stenosis with and without coexisting spondylolisthesis and scoliosis

Implementing machine learning techniques, such as decision trees, known as prediction models that use logical construction diagrams, are rarely used to predict clinical outcomes. To develop a clinical prediction rule to predict clinical outcomes in patients who undergo minimally invasive lumbar decompression surgery for lumbar spinal stenosis with and without coexisting spondylolisthesis and scoliosis using a decision tree model. A retrospective analysis of prospectively collected data. This study included 331 patients who underwent minimally invasive surgery for lumbar spinal stenosis and were followed up for ≥2 years at 1 institution. Self-report measures: The Japanese Orthopedic Association (JOA) scores and low back pain (LBP)/leg pain/leg numbness visual analog scale (VAS) scores. Physiologic measures: Standing sagittal spinopelvic alignment, computed tomography, and magnetic resonance imaging results. Low achievement in clinical outcomes were defined as the postoperative JOA score at the 2-year follow-up <25 points. Univariate and multiple logistic regression analysis and chi-square automatic interaction detection (CHAID) were used for analysis. The CHAID model for JOA score <25 points showed spontaneous numbness/pain as the first decision node. For the presence of spontaneous numbness/pain, sagittal vertical axis ≥70 mm was selected as the second decision node. Then lateral wedging, ≥6° and pelvic incidence minus lumbar lordosis (PI-LL) ≥30° followed as the third decision node. For the absence of spontaneous numbness/pain, sex and lateral olisthesis, ≥3mm and American Society of Anesthesiologists physical status classification system score were selected as the second and third decision nodes. The sensitivity, specificity, and the positive predictive value of this CHAID model was 65.1, 69.8, and 64.7% respectively. The CHAID model incorporating basic information and functional and radiologic factors is useful for predicting surgical outcomes.

A simple algorithm to predict disability in community-dwelling older Japanese adults

BackgroundWith a worldwide aging population, the prevention of disability in older adults has become an important issue. Therefore, the purpose of this study was to develop a model for predicting disability risk in older adults based on multiple factors, using a decision tree analysis. This model may be used with a mobile application when it is difficult to interview older adults, and to obtain individualized information for prioritizing interventions. MethodsWe examined the data from a cohort study conducted by the National Center for Geriatrics and Gerontology-Study of Geriatric Syndromes. We included 12,000 older adults without a disability and performed a decision tree analysis using the Chi-square automatic interaction detection (CHAID) algorithm. ResultsAmong the 12,000 participants without a disability, 11,503 and 497 participants remained disability-free and developed disability, respectively. The CHAID analysis identified 24 end nodes with five levels of partition and 16 partitioning variables for 34 questionnaire variables, with incident disability probabilities ranging from 0.0% to 96.7%. The classification accuracy and area under the curve of the CHAID model were 73.4% and 0.76, respectively. We found that maintaining mental health was important for older adults in their 80s and older, and that lifestyles and geriatric syndromes were important factors for those in their 70s. ConclusionsThe magnitude of the influences on the risk of developing a disability differ by age group. The results of this study may provide useful information for the development of mobile applications that predict the risk of developing disability and create tailor-made interventions.

Exploring the predictive capability of machine learning models in identifying foot and mouth disease outbreak occurrences in cattle farms in an endemic setting of Thailand

Occurrences of foot and mouth disease (FMD) outbreaks in cattle farms in Thailand have been significantly harmful to the cattle industry for the past decade. A prediction of FMD outbreaks based on relevant risk factors with a high prediction accuracy is important for authorities to develop a plan for preventing the outbreaks. Data-driven tools are widely accepted for their prediction abilities, but an application of these techniques to FMD outbreak prediction is very limited. The objectives of this study were to develop prediction models of FMD outbreaks among cattle farms using machine learning (ML) classification algorithms including classification tree (CT), random forests (RF), and Chi-squared automatic interaction detection (CHAID) and to compare the predictive performance of the developed models. Data from 225 FMD and 608 non-FMD outbreak farms from an endemic setting were analyzed using ML methods. The CT, RF, and CHAID methods were utilized to develop predictive models, and their prediction capabilities were compared. The results showed that models developed using ML methods have an acceptable to excellent ability to predict the occurrence of FMD outbreaks. The RF model had the highest accuracy and the value of area under the operating characteristic curve in predicting the occurrence of an FMD outbreak. Meanwhile, the CT and CHAID models delivered comparable results. In this study, we demonstrated the capability of machine learning algorithms to predict FMD outbreaks using actual FMD outbreak data from the endemic setting and provided a new insight into the prediction of FMD outbreaks. The ML techniques demonstrated herein may be used as a prediction tool by the relevant authorities to predict the occurrence of FMD outbreaks in cattle farms.

Detection of subclinical keratoconus using a novel combined tomographic and biomechanical model based on an automated decision tree

Early detection of keratoconus is a crucial factor in monitoring its progression and making the decision to perform refractive surgery. The aim of this study was to use the decision tree technique in the classification and prediction of subclinical keratoconus (SKC). A total of 194 eyes (including 105 normal eyes and 89 with SKC) were included in the double-center retrospective study. Data were separately used for training and validation databases. The baseline variables were derived from tomography and biomechanical imaging. The decision tree models were generated using Chi-square automatic interaction detection (CHAID) and classification and regression tree (CART) algorithms based on the training database. The discriminating rules of the CART model selected metrics of the Belin/Ambrósio deviation (BAD-D), stiffness parameter at first applanation (SPA1), back eccentricity (Becc), and maximum pachymetric progression index in that order; On the other hand, the CHAID model selected BAD-D, deformation amplitude ratio, SPA1, and Becc. Further, the CART model allowed for discrimination between normal and SKC eyes with 92.2% accuracy, which was higher than that of the CHAID model (88.3%), BAD-D (82.0%), Corvis biomechanical index (CBI, 77.3%), and tomographic and biomechanical index (TBI, 78.1%). The discriminating performance of the CART model was validated with 92.4% accuracy, while the CHAID model was validated with 86.4% accuracy in the validation database. Thus, the CART model using tomography and biomechanical imaging was an excellent model for SKC screening and provided easy-to-understand discriminating rules.

Developing Machine Learning Techniques to Investigate the Impact of Air Quality Indices on Tadawul Exchange Index

The air quality index (AQI) can be described using different pollutant indices. Many investigators study the effect of stock prices and air quality in the field to show if there is a relationship between changing the stock market and the concentration of various pollutants. This study aims to find a relationship between Saudi Tadawul All Share Index (TASI) and multiple pollutants, including particulate matter (PM10), ozone (O3), nitrogen dioxide (NO2), sulfur dioxide (SO2), carbon monoxide (CO), and AQI. Based on tree models, the relationship is created using linear regression and two prediction models, Chi‐square Automatic Interaction Detection (CHAID), and CR‐Tree. In order to achieve the target of this research, the TASI dataset relates to six pollutants using time; afterward, the new dataset is divided into three parts—test, validate, and train—after eliminating the outlier data. In order to test the performance of two prediction models, R2 and various error functions are used. The linear regression model results found that PM10, NO2, CO, month, day, and year are significant, whereas O3, SO2, and AQI indices are insignificant. The test dataset showed that R2 scores are 0.995 and 0.986 for CR‐Tree and CHAID, respectively, with RMSE values of 387 and 227 for CR‐Tree and CHAID, respectively. The prediction models showed that the CHAID model performed better than CR‐Tree because it used only three indices, namely, PM10, AQI, and O3, with year and month. The results indicated an effect between changing TASI and the three pollutants, PM10, AQI, and O3.

Decision tree analysis to predict the risk of intracranial haemorrhage after mild traumatic brain injury in patients taking DOACs.

Although the preliminary evidence seems to confirm a lower incidence of post-traumatic bleeding in patients treated with direct oral anticoagulants (DOACs) compared to those on vitamin K antagonists (VKAs), the recommended management of mild traumatic brain injury (MTBI) in patients on DOACs is the same as those on the older VKAs, risking excessive use of CT in the emergency department (ED). To determine which easily identifiable clinical risk factors at the first medical evaluation in the ED may indicate an increased risk of post-traumatic intracranial haemorrhage (ICH) in patients on DOACs with MTBI. Patients on DOACs who were evaluated in the ED for an MTBI from 2016 to 2020 at four centres in Northern Italy were considered. A decision tree analysis using the chi-square automatic interaction detection (CHAID) method was conducted to assess the risk of post-traumatic ICH after an MTBI. Known pre- and post-traumatic clinical risk factors that are easily identifiable at the first medical evaluation in the ED were used as input predictor variables. Among the 1146 patients on DOACs in this study, post-traumatic ICH was present in 6.5% (75/1146). Decision tree analysis using the CHAID method found post-traumatic TLOC, post-traumatic amnesia, major trauma dynamic, previous neurosurgery and evidence of trauma above the clavicles to be the strongest predictors associated with the presence of post-traumatic ICH in patients on DOACs. The absence of a concussion seems to indicate subgroups at very low risk of requiring neurosurgery. The machine-based CHAID model identified distinct prognostic groups of patients with distinct outcomes based on clinical factors. Decision trees can be useful as guides for patient selection and risk stratification.

Exploring Quality of Life Reported by Norwegian Older Adults Using Classification Tree Approach on Group Profiles

This study aims to explore the variation between- and within subgroups of older adults with regard to low, medium, and high levels of self-reported quality of life (QoL) measured by the WHOQOL-BREF scale. The contribution of interacting personal and contextual life conditions to QoL was examined in a sample of 1,910 (sample frame 6,000) Norwegian men and women aged 62 to 99 years. The data collected by a postal questionnaire were analyzed using the Chi-square Automatic Interaction Detection (CHAID) classification method in order to detect unique profiles of groups who shared common characteristics. The CHAID model revealed 15 relatively homogenous groups, but distinct from one another, whose profiles were defined by unique constellations of several interacting variables significantly related to a given QoL level. Mental functioning was predominantly linked to perception of life meaning along with health status, and/or in some cases living arrangement, loneliness, neighborhood quality, and satisfaction with income, and placed an individual at different likelihood levels of reporting low, medium, or high QoL. Socio-demographics had no statistically significant impact on QoL for any subgroup. Through this individual-oriented approach, a periodically ongoing assessment of subjective quality of life (QoL) may be sufficiently powerful to allow detecting and addressing personal concerns and specific needs that detract from quality of life in advancing age.

Machine Learning Algorithm Identifies the Importance of Environmental Factors for Hospital Discharge to Home of Stroke Patients using Wheelchair after Discharge

Background and purposePhysical environmental factors are generally likely to become barriers for discharge to home of wheelchair users, compared with non-wheelchair users. However, the importance of environmental factors has not been investigated adequately. Application of machine learning technology might efficiently identify the most influential factors, although it is not easy to interpret and integrate various information including individual and environmental factors in clinical stroke rehabilitation. This study aimed to identify the influential factors affecting home discharge in the stroke patients who use a wheelchair after discharge by using machine learning technology. MethodsThis study used the rehabilitation database of our facility, which includes all stroke patients admitted into the convalescence rehabilitation ward. The chi-squared automatic interaction detection (CHAID) algorithm was used to develop a model to classify wheelchair-using stroke patients discharged to home or not-to-home. ResultsAmong the variables, including basic information, motor functional factor, activities of daily living ability factor, and environmental factors, the CHAID model identified house renovation and the existence of sloping roads around the house as the first and second discriminators for home discharge. ConclusionsOur present results could scientifically clarify that the clinician need to focus on the physical environmental factors for achieving home discharge in the patients who use a wheelchair after discharge.

Classification of transformed anchovy products based on the use of element patterns and decision trees to assess traceability and country of origin labelling

Quadrupole inductively coupled plasma mass spectrometry (Q-ICP-MS) and direct mercury analysis were used to determine the elemental composition of 180 transformed (salt-ripened) anchovies from three different fishing areas before and after packaging. To this purpose, four decision trees-based algorithms, corresponding to C5.0, classification and regression trees (CART), chi-squareautomatic interaction detection (CHAID), and quick unbiased efficient statistical tree (QUEST) were applied to the elemental datasets to find the most accurate data mining procedure to achieve the ultimate goal of fish origin prediction. Classification rules generated by the trained CHAID model optimally identified unlabelled testing bulk anchovies (93.9% F-score) by using just 6 out of 52 elements (As, K, P, Cd, Li, and Sr). The finished packaged product was better modelled by the QUEST algorithm which recognised the origin of anchovies with F-score of 97.7%, considering the information carried out by 5 elements (B, As, K. Cd, and Pd). Results obtained suggested that the traceability system in the fishery sector may be supported by simplified machine learning techniques applied to a limited but effective number of inorganic predictors of origin.

Clinical Features for Identifying the Possibility of Toileting Independence after Convalescent Inpatient Rehabilitation in Severe Stroke Patients: A Decision Tree Analysis Based on a Nationwide Japan Rehabilitation Database

Background and purposeIn severe stroke patients, considerable concern should be given to toileting activity in rehabilitative support. Recently, the application of artificial intelligence, including machine learning (ML), has expanded into the stroke medical field, which could clarify the factors affecting toileting independence in severe stroke patients. This study aimed to identify the factors affecting toileting independence in severe stroke patients using ML. MethodsWe used the Japan Rehabilitation Database from 2005 to 2015 to investigate data from 2292 severe stroke patients. We performed the chi-squared automatic interaction detection (CHAID) algorithm with various explanatory variables. ResultsThe CHAID model identified modified Rankin scale (mRS) score as the first discriminator. Among those with an mRS score ≤4, the next discriminator was age (score ≤72, 73–80, or >80). Among those with an mRS score > 4, the next discriminator was also age (score ≤57, 58–72, 73–80, or >80). Interestingly, some patients achieved toileting independence, although this study focused on severe stroke patients. In branches based on age, the percentage of the patients who achieved toileting independence at discharge decreased progressively with age. ConclusionWe identified the influential factors, including reference values, for achieving toileting independence in convalescent severe stroke patients.

Classification-Based Regression Models for Prediction of the Mechanical Properties of Roller-Compacted Concrete Pavement

In the field of pavement engineering, the determination of the mechanical characteristics is one of the essential processes for reliable material design and highway sustainability. Early determination of the mechanical characteristics of pavement is essential for road and highway construction and maintenance. Tensile strength (TS), compressive strength (CS), and flexural strength (FS) of roller-compacted concrete pavement (RCCP) are crucial characteristics. In this research, the classification-based regression models random forest (RF), M5rule model tree (M5rule), M5prime model tree (M5p), and chi-square automatic interaction detection (CHAID) are used for simulation of the mechanical characteristics of RCCP. A comprehensive and reliable dataset comprising 621, 326, and 290 data records for CS, TS, and FS experimental cases was extracted from several open sources in the literature. The mechanical properties are determined based on influential input combinations that are processed using principle component analysis (PCA). The PCA method specifies that volumetric/weighted content forms of experimental variables (e.g., coarse aggregate, fine aggregate, supplementary cementitious materials, water, and binder) and specimens’ age are the most effective inputs to generate better performance. Several statistical metrics were used to evaluate the proposed classification-based regression models. The RF model revealed an optimistic classification capacity of the CS, TS, and FS prediction of the RCCP in comparison with the CHAID, M5rule, and M5p models. Monte-Carlo simulation was used to verify the results in terms of the uncertainty and sensitivity of variables. Overall, the proposed methodology formed a reliable soft computing model that can be implemented for material engineering, construction, and design.

Landslide Susceptibility Prediction Based on Remote Sensing Images and GIS: Comparisons of Supervised and Unsupervised Machine Learning Models

Landslide susceptibility prediction (LSP) has been widely and effectively implemented by machine learning (ML) models based on remote sensing (RS) images and Geographic Information System (GIS). However, comparisons of the applications of ML models for LSP from the perspectives of supervised machine learning (SML) and unsupervised machine learning (USML) have not been explored. Hence, this study aims to compare the LSP performance of these SML and USML models, thus further to explore the advantages and disadvantages of these ML models and to realize a more accurate and reliable LSP result. Two representative SML models (support vector machine (SVM) and CHi-squared Automatic Interaction Detection (CHAID)) and two representative USML models (K-means and Kohonen models) are respectively used to scientifically predict the landslide susceptibility indexes, and then these prediction results are discussed. Ningdu County with 446 recorded landslides obtained through field investigations is introduced as case study. A total of 12 conditioning factors are obtained through procession of Landsat TM 8 images and high-resolution aerial images, topographical and hydrological spatial analysis of Digital Elevation Modeling in GIS software, and government reports. The area value under the curve of receiver operating features (AUC) is applied for evaluating the prediction accuracy of SML models, and the frequency ratio (FR) accuracy is then introduced to compare the remarkable prediction performance differences between SML and USML models. Overall, the receiver operation curve (ROC) results show that the AUC of the SVM is 0.892 and is slightly greater than the AUC of the CHAID model (0.872). The FR accuracy results show that the SVM model has the highest accuracy for LSP (77.80%), followed by the CHAID model (74.50%), the Kohonen model (72.8%) and the K-means model (69.7%), which indicates that the SML models can reach considerably better prediction capability than the USML models. It can be concluded that selecting recorded landslides as prior knowledge to train and test the LSP models is the key reason for the higher prediction accuracy of the SML models, while the lack of a priori knowledge and target guidance is an important reason for the low LSP accuracy of the USML models. Nevertheless, the USML models can also be used to implement LSP due to their advantages of efficient modeling processes, dimensionality reduction and strong scalability.