Polynomial-Fuzzy Decision Tree Structures for Classifying Medical Data

In this paper, we present a novel ensemble method random projection random discretization ensembles (RPRDE) to create ensembles of linear multivariate decision trees by using a univariate decision tree algorithm. The present method combines the better computational complexity of a univariate decision tree algorithm with the better representational power of linear multivariate decision trees. We develop random discretization (RD) method that creates random discretized features from continuous features. Random projection (RP) is used to create new features that are linear combinations of original features. A new dataset is created by augmenting discretized features (created by using RD) with features created by using RP. Each decision tree of a RPRD ensemble is trained on one dataset from the pool of these datasets by using a univariate decision tree algorithm. As these multivariate decision trees (because of features created by RP) have more representational power than univariate decision trees, we expect accurate decision trees in the ensemble. Diverse training datasets ensure diverse decision trees in the ensemble. We study the performance of RPRDE against other popular ensemble techniques using C4.5 tree as the base classifier. RPRDE matches or outperforms other popular ensemble methods. Experiments results also suggest that the proposed method is quite robust to the class noise.

To develop and test decision tree (DT) models to classify physical activity (PA) intensity from accelerometer output and Gross Motor Function Classification System (GMFCS) classification level in ambulatory youth with cerebral palsy (CP) and compare the classification accuracy of the new DT models to that achieved by previously published cut points for youth with CP. Youth with CP (GMFCS levels I-III) (N = 51) completed seven activity trials with increasing PA intensity while wearing a portable metabolic system and ActiGraph GT3X accelerometers. DT models were used to identify vertical axis (VA) and vector magnitude (VM) count thresholds corresponding to sedentary (SED) (<1.5 METs), light-intensity PA (LPA) (≥1.5 and <3 METs) and moderate-to-vigorous PA (MVPA) (≥3 METs). Models were trained and cross-validated using the "rpart" and "caret" packages within R. For the VA (VA_DT) and VM DT (VM_DT), a single threshold differentiated LPA from SED, whereas the threshold for differentiating MVPA from LPA decreased as the level of impairment increased. The average cross-validation accuracies for the VC_DT were 81.1%, 76.7%, and 82.9% for GMFCS levels I, II, and III. The corresponding cross-validation accuracies for the VM_DT were 80.5%, 75.6%, and 84.2%. Within each GMFCS level, the DT models achieved better PA intensity recognition than previously published cut points. The accuracy differential was greatest among GMFCS level III participants, in whom the previously published cut points misclassified 40% of the MVPA activity trials. The GMFCS-specific cut points provide more accurate assessments of MVPA levels in youth with CP across the full spectrum of ambulatory ability.

Health economics is a discipline of economics applied to health care. One method used in health economics is decision tree modelling, which extrapolates the cost and effectiveness of competing interventions over time. Such decision tree models are the basis of reimbursement decisions in countries using health technology assessment for decision making. In many instances, these competing interventions are diagnostic technologies. Despite a wealth of excellent resources describing the decision analysis of diagnostics, two critical errors persist: not including diagnostic test accuracy in the structure of decision trees and treating sequential diagnostics as independent. These errors have consequences for the accuracy of model results, and thereby impact on decision making. This paper sets out to overcome these errors using color to link fundamental epidemiological calculations to decision tree models in a visually and intuitively appealing pictorial format. The paper is a must-read for modelers developing decision trees in the area of diagnostics for the first time and decision makers reviewing diagnostic reimbursement models.

Teicoplanin 24-h loading dose regimen using a decision tree model to target serum trough concentration of 15–30 μg/mL: A retrospective study

Multiple machine learning (ML) algorithms were developed using artificial intelligence, including Linear Regression (LR), Decision Tree (DT), Random Forest (RF), and K-Nearest Neighbor (KNN), to predict the yield of biodiesel production in an acid-catalyzed alcoholysis process using green seed canola oil. Catalyst loading, methanol-to-oil (M/O) molar ratio, and reaction time were considered as input parameters, while the yield of biodiesel production was selected as the output parameter. The performance of the developed ML models was assessed using evaluation metrics such as the coefficient of determination (R2) and the root mean squared error (RMSE). The R2 values obtained for LR, RF, DT, and KNN models were 0.80, 0.95, 0.97, and 0.84, respectively. Furthermore, the corresponding RMSE values for these models were 2.48, 1.51, 0.89, and 4.51, respectively. According to the results, the DT model exhibited superior accuracy and reliability for predicting biodiesel production compared to the other models. The values of the input variables to potentially yield the highest biodiesel output were identified through a systematic trial-and-error approach using the DT model. The results showed that a biodiesel yield of 88 % can be achieved with 5 wt% catalyst loading, a 22 M/O molar ratio, and a reaction time of 5 hours.

The analyzed study focuses on the application of predictive models, specifically Linear Regression and Decision Trees, for the management of delinquent debts in the public context of the United States. The main objective of the work is to compare the effectiveness of these models in predicting the compliance of debts older than 120 days, helping to direct these debts to the Treasury Offset Program (TOP), an essential initiative for government financial recovery. The problem that the study addresses is the need for effective management of delinquent public debts, seeking to ensure compliance with public financial policies that promote compliance and the proper redirection of financial resources to the government. This is particularly important to ensure fiscal transparency and accountability of federal agencies. The methodology used in the study was quantitative, based on the analysis of eligible debt data extracted from U.S. Treasury reports. The Linear Regression and Decision Trees models were applied, with performance metrics such as Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE) and Coefficient of Determination (R²). The study dealt with financial and temporal variables to analyze the behavior of these debts and their compliance. The main results show that both models showed high accuracy in the predictions, with the Linear Regression showing a perfect fit (R² = 1) and the Decision Trees excelling in capturing nonlinear nuances of the data. The "Compliance Rate Amount" variable was identified as the most significant in the Decision Tree model, suggesting that the amount of the compliance rate is one of the most important factors to predict the compliance of delinquent debts. This study offers valuable contributions to the field of public management, by demonstrating that the use of predictive models can help optimize debt recovery, improve fiscal transparency, and contribute to more informed decision-making.

The purpose of the work is to investigate the use of decision tree (DT) enhanced by bootstrap aggregates (Bag) and least-squares boosting (Lsboost) in modeling the organic matter of water according to its physicochemical parameters. An entire database of 500 samples of 21 physicochemical parameters, including organic matter, was used to build the DT, DT_Bag, and DT_Lsboost models. Training data (364 data points) is resampled using a bootstrap technique to form different training datasets to train different models. The models built were validated by a dataset of 91 samples. The predicted outputs obtained from the developed DT models are then combined by simple averaging. On the other hand, the data was also boosted with the Lsboost technical aid to increase the strength of a weak learning algorithm. The model trains the first weak learner with equal weight across all data points in the training set, then trains all other weak learners based on the updated weight aimed at the validation result to minimize the squared error medium. Good agreement between the predicted and experimental organic matter concentrations for the DT_Lsboost model was obtained (the correlation coefficient for the validation dataset was 0.9992), followed by the DT-Bag model with a correlation coefficient of 0.9949. The comparison between DT, DT_Bag, and DT_Lsboost revealed the superiority of the DT_Lsboost model (the mean root of the squared errors for the dataset were 0.1295 for the DT_Lsboost, 0.1664 for the DT_Bag, and 0.5444 for the DT). These results show that Lsboost technology dramatically improved the DT model. This result is also confirmed by the results of tests on models (interpolation data of 45 points). It should also be noted that the Bag technique was also very effective in optimizing the DT model, as the results obtained with this technique were very close to the DT_Lsboost model.

ObjectiveThis study investigates the factors influencing sedentary behavior in older adult Chinese stroke patients using decision trees and logistic regression models.MethodsConvenience sampling method was employed to enroll 346 respondents aged ≥60 years with stroke from the Department of Neurology of three tertiary-level A hospitals in Heilongjiang province, based on the inclusion criteria. The Sedentary Behavior Questionnaire for Older Adults, the International Physical Activity Questionnaire Short Form (IPAQ-S), the Pittsburgh Sleep Quality Index (PSQI), the Self-Rating Depression Scale (SDS), and the Social Support Scale (SSRS) were used to assess sedentary behavior, physical activity level, sleep quality, depressive symptoms, and social support, respectively. Decision tree and logistic regression models were employed to analyze the factors related to sedentary behavior in older adult stroke patients.ResultsOf the 346 respondents, 233 (67.3%) had sedentary behavior. The logistic regression model showed that education level (OR = 2.843, 95%CI: 1.219–6.626), BMI (OR = 3.686, 95%CI: 1.838–7.393), longest consecutive sitting time (OR = 3.853, 95%CI: 1.867–7.953), and sleep quality (OR = 3.832, 95%CI: 1.716–8.557) were identified as risk factors for sedentary behavior in older adult stroke patients, while drink alcohol (OR = 0.386, 95%CI: 0.184–0.809) and physical activity level (OR = 0.064, 95%CI: 0.030–0.140) were identified as protective factors for sedentary behavior. Besides, the decision tree model showed that physical activity level, longest consecutive sitting time, sleep quality, BMI, depressive symptoms, and age were associated with sedentary behavior. The sensitivity and specificity of the logistic regression model were 69.9 and 93.1%, respectively, and the area under the receiver operating characteristic (ROC) curve was 0.900 (95% CI: 0.863–0.938). The sensitivity and specificity of the decision tree model were 66.4, and 93.1% respectively, and the area under the ROC curve was 0.860 (95% CI: 0.816–0.904).ConclusionOur findings indicated that physical activity level, longest consecutive sitting time, sleep quality, and BMI were key factors associated with sedentary behavior. To achieve the purpose of improving rehabilitation effect and quality of life, this study combining decision trees with logistic regression models was of high value in studying factors influencing sedentary behavior in older adult stroke patients.

Several publications concerning decision tree (DT) analysis in medical fields have recently demonstrated its usefulness for defining prognostic factors in various diseases. However, there are minimal reports on the predictors of adverse drug reactions. We attempted to use DT analysis to discover combinations of multiple risk factors that would increase the risk of nephrotoxicity associated with vancomycin (VCM). To demonstrate the usefulness of DT analysis, we compared its predictive performance with that of multiple logistic regression analysis. A single-centre, retrospective study was conducted at Hokkaido University Hospital. A total of 592 patients, who received intravenous administrations of VCM between November 2011 and April 2016, were enrolled. Nephrotoxicity was defined as an increase in serum creatinine of ≥0.5mg/dL or a ≥50% increase in serum creatinine from the baseline. Risk factors for VCM nephrotoxicity were extracted from previous reports. In the DT analysis, a chi-squared automatic interaction detection algorithm was constructed. For evaluating the established algorithms, a 10-fold cross validation method was adopted to calculate the misclassification risk of the model. Moreover, to compare the accuracy of the DT analysis, multiple logistic regression analysis was conducted. Eighty-seven (14.7%) patients developed nephrotoxicity. A VCM trough concentration of ≥15.0mg/L, concomitant medication (vasopressor drugs and furosemide), and a duration of therapy ≥14days were extracted to build the DT model, in which the patients were divided into 6 subgroups based on variable rates of nephrotoxicity, ranging from 4.6 to 69.6%. The predictive accuracies of the DT and logistic regression models were similar (87.3%, respectively), indicating that they were accurate. This study suggests the usefulness of DT models for the evaluation of adverse drug reactions.

Badminton is a demanding sport that requires effective workload management to enhance performance and prevent injuries. This study developed a machine learning-based Decision Tree (DT) model to create personalized workload management strategies for 73 young elite badminton players, averaging 6 years of experience. Players underwent anthropometric and fitness assessments, with external loads measured via triaxial accelerometers and internal loads through rate of perceived exertion (RPE) during training and competition. K-means clustering categorized players into high, moderate, and low external workload levels. High-load players were generally older, taller, heavier, and exhibited superior flexibility, grip strength, and countermovement jump performance. Moderate-load players excelled in balance and leg endurance, while low-load players showed greater upper body strength, quicker reaction times, and higher perceived exertion. A sensitivity analysis was conducted to evaluate the impact of tree depth on model performance, followed by a comparative assessment of the Decision Tree (DT) model and multinomial Logistic Regression (MLR). The results demonstrated that the DT model outperformed the MLR, achieving 92% accuracy in predicting external loads compared to the MLR's 57%. This highlights the DT model's superior capability to provide tailored workload recommendations, thereby enhancing athletic performance and reducing the risk of injury.

ObjectiveThis study aimed to systematically identify risk factors for urinary catheter-related hematuria in patients with acute myocardial infarction (AMI). By integrating logistic regression and decision tree models, we sought to develop actionable strategies for risk stratification and complication prevention.MethodsA retrospective analysis of 209 AMI patients was conducted to evaluate predictors of hematuria, including demographics, coagulation indices (INR, platelets), and procedural variables. Logistic regression and decision tree (CART algorithm) models were employed to identify risk factors and their interactions. Model performance was assessed using ROC-AUC, sensitivity, and specificity.ResultsThe incidence of catheter-related hematuria was 32.5%. Both models identified persistent agitation during catheter indwelling and PLT ≤ 246 as common predictors. The logistic regression model specifically identified Gender (OR = 0.202), patient awareness of catheter purpose and precautions (OR = 0.470), and emergency catheter placement (OR = 2.257) as significant factors. The decision tree model uniquely identified INR > 0.955 and repeated complaints of urethral pain as predictors.ConclusionHematuria in AMI patients results from coagulation dysfunction, procedural trauma, and behavioral factors. The combined use of logistic regression and decision trees enhances risk stratification. Clinical strategies should prioritize gentle catheterization, dynamic coagulation monitoring, and patient education to reduce complications.

Regional landslide susceptibility prediction (LSP) research is of great significance to the prevention and control of landslides. This study focuses on the LSP modelling based on the decision tree model. Taking the northern part of An’yuan County of Jiangxi Province as an example, 14 environmental factors including elevation, gully density and lithology are obtained based on geographical information system (GIS) and remote sensing satellite. Frequency Ratio method and C5.0 decision tree (DT) model are coupled to build DT model for LSP modelling. Then the predicted results are graded into five attribute intervals. Finally, LSP performance of DT model is evaluated by comparing the area value under the receiver operating characteristic curve (ROC) and classification of landslide susceptibility. The results show that the AUC accuracy of the C5.0 DT model is 0.805, and the LSP results of the C5.0 DT model are consistent with the actual distribution pattern of landslides in this County.

To study the application of decision tree in the research of anemia among rural children. In the Enterprise Miner module of software SAS 8.2, 3000 observations were sampled from database and the decision tree model was built. The model using decision tree of CART bases on Gini impurity index. The misclassification rate of decision tree model was, training set 21.2%, validation set 21.9%. The Root ASE of decision tree model was, training set 0.399, validation set 0.404. The area under the ROC curve was larger than the reference line. The diagnostic chart showed that the corresponding percentage was higher than the other. The decision tree model selected 9 important factors and ranked them by their power, among which mother of anemia (1.00) was the most important factor. Others were children's age (0.75), time of ablactation (0.53), mother's age (0.32), the time of egg supplementation (0.26), category of the project county (0.26), the time of milk supplementation (0.16), number of people in the family (0.13), the education status of the mother (0.12). Decision tree produced simple and easy rules that might be used to classify and predict in the same research. Decision tree could screen out the important factors of anemia and identify the cutting-points for factors. With the wide application of decision tree, it would exhibit important application values in the research of the rural children health care.

Statistical models are essential to understand the occurrence and spread of microbes to support decision-making in microbiology and epidemiology. Antimicrobial resistance (AMR) is a multifaceted global problem and a significant threat to sustainable modern healthcare. This thesis aims to identify knowledge gaps in the AMR research field and explain the added value of using statistical models and novel spatiotemporal data to predict and identify risk factors for the occurrence and spread of AMR. Strategic action plans to tackle the increasing international threat of AMR are based upon research agendas that are informed using knowledge gaps in the AMR research field. Currently, these knowledge gaps are identified manually and are often subjective. Chapter 2 describes how bibliometric data-driven methodology can be used to identify knowledge gaps in AMR research. To this end, twenty years of AMR related articles were extracted using the PubMed search engine. With structural topic modelling I identified the topics comprising the AMR research field, while topic clusters were created using hierarchical clustering on the topic proportions. Potential AMR knowledge gaps were obtained using Spearman’s correlation between topic clusters and topics and between individual topics. A total of 88 topics and seven topic clusters were identified from 158 616 scientific AMR research articles. In total, 421 potential knowledge gaps were identified between the topic clusters and topics and 2 663 between individual topics. Key knowledge gaps between molecular and laboratory AMR research were highlighted. The knowledge gaps between AMR research regarding water and the environment and both institutional and international surveillance topics were highlighted at the topic level. These results provide an innovative, data-driven way to identify knowledge gaps in AMR research. Surgical site infections (SSI) make up 19.6% of healthcare-associated infections (HAIs) in Europe [98]. Risk factor identification studies for the occurrence of SSI do not usually specify how continuous variables cut-offs are determined. In most cases, they use standard medical cut-offs without considering the data being studied. Chapter 3 identifies the risk factors for the occurrence of SSI for digestive, thoracic and orthopaedic system surgeries using standard medical and data-driven cut-off values. Retrospective surgical procedure data, individual electronic health records, pharmaceutical data and laboratory data were used from the Erasmus MC University Medical Centre in The Netherlands. Risk factors for the occurrence of SSI were identified using a multivariate forward-step logistic regression model. Standard medical cut-off values were compared with cut-offs determined from the data. For digestive, orthopaedic and thoracic system surgical procedures, the risk factors identified for the occurrence of SSI were preoperative temperature of 38 oC and antibiotics used at the time of surgery. C-reactive protein (CRP) and the duration of the surgery were identified as risk factors for digestive surgical procedures. Being an adult (age ≥ 18) was identified as a protective effect for thoracic surgical procedures. Data-driven cut-off values identified for temperature, age, and CRP, explained the occurrence of SSI outcome up to 19.5% better than standard medical cut-off values. Future studies should investigate if data-driven cut-offs can add value to explain the clinical outcome being modelled and not solely rely on standard medical cut-off values for continuous variables to identify risk factors. Transmission of harmful microorganisms (HMO) poses a major threat to patients and healthcare workers in healthcare settings. The most effective countermeasure against these transmissions is the adherence to hand hygiene policies, but adherence rates are relatively low and vary over space and time. The spatiotemporal effects of varying levels hand hygiene compliance on the transmission and spread of hand-transmitted HMO in a closed healthcare setting must still be quantified. Chapter 4 describes how identifies healthcare worker occupation group of potential super-spreaders and the spatiotemporal effects on the hand transmission of HMO quantified for varying levels of hand hygiene compliance (HHC) caused by this group using their spatiotemporal movements. Spatiotemporal data were collected in the University Medical Center Groningen (UMCG) using radio frequency identification technology. The effects of five probability distributions of HHC and three harmful microorganism transmission rates were simulated using a dynamic agent-based simulation model. The effects of initial simulation assumptions on the simulation results were quantified using five risk outcomes. Nurses were identified as the potential super-spreader healthcare worker occupation group. During lack of HHC (5%) and high transmission rates (5% per contact moment), a colonised nurse can transfer microbes to three of the 17 healthcare worker or patients encountered during the 98.4 minutes of visiting 23 rooms while colonised. The HMO transmission potential for nurses is higher during weeknights (5 pm – 7 am) and weekends as compared to weekdays (7 am – 5 pm). Spatiotemporal behaviour and social mixing patterns of healthcare can change the expected number of hand transmissions and spread HMO by super-spreaders in a closed healthcare setting. These insights can be used to evaluate spatiotemporal safety behaviours and develop infection prevention and control strategies. Vancomycin-resistant enterococci (VRE) is can cause severe patient health and monetary burdens. The odds of a hospital patient acquiring VRE increases when using antibiotics and when prior room occupants had VRE, but the antibiotic use of prior room occupants are often neglected. Chapter 5 describes how the occurrence and spread of VRE can be explained using intrahospital patient movements (IPM) and their antibiotic use between hospital wards. Retrospective IPM, antibiotic use and PCR screening data were used from a hospital in the Netherlands. A dynamic directed spatiotemporal graph was developed, and together with the PageRank algorithm used to calculate two daily centrality measures to summarise the flow of patients and antibiotics at the ward level. With a decision tree and random forest model I predicted the daily occurrence of VRE for every ward and compared the models’ performance using a 30% test sample. The decision tree model produced a simple set of rules that can be used to determine the daily probability of VRE occurrence for each hospital ward. The decision tree model achieved an area under the curve of 0.685 and the random forest model 0.886 on the test set. These results confirm that the random forest model performs better than a single decision tree for all levels of model sensitivity and specificity at the cost of model simplicity. An early warning system for VRE can be developed and inform infection prevention plans and outbreak strategies further using these results. In summary, this thesis showed that data-driven statistical models can improve our understanding of antimicrobial resistance. It considers how different sources of spatiotemporal data may be used to predict its occurrence and spread of AMR in hospitals.

The classification model in machine learning has been employed to address different problems. Machine Learning classification is an effective method to realize customer churn prediction. This article provides a comparative study of machine learning from the perspective of predicting customer churn. Customer churn is one of the obstacles hindering the development of companies. Through classification approaches based on machine learning, customer churn can be predicted precisely, thus providing decision-making capabilities to these companies. Customer churn prediction is a typical classification problem and can be addressed by the use of a decision tree or random forest model. In this paper, decision tree and random forest models are employed to predict customer churn, using sales data provided by a chemical company from 2012 to 2020. We analyze the underlying risk of customer churn and the customer churn factors, including the low-priced count (LC), total amount of money (TM), and creation time (CT). The prediction results of the two models are evaluated by calculating various metrics. The experimental results indicate that the low-priced count (LC) is the most essential factor for customer churn. The results of the comparison, considering the training error and generalization error including: confusion matrix, ROC curve, AUC, precision, recall, and F1 score, reveal that the random forest model has better prediction accuracy than the decision tree model.