Decision Tree Model For Prediction Research Articles

Epidemiology, Seasonal Variability, and Factors Associated with Increased Mortality in Patients with Acute Appendicitis in Thailand: A Study Using Data from the National Health Security Office

Objective: The study aimed to study the spatiotemporal distribution of acute appendicitis in Thailand, using a reimbursement dataset from the National Health Security Office, and also analyze for mortality and factors associated with mortality after an appendectomy.Material and Methods: The study used data mining to analyze a reimbursement data set from the Thailand National Health Security Office (NHSO), focusing on the diagnosis of acute appendicitis (K35.2-K35.9). The analysis looked for the incidence trend and seasonal variation of the incidence. The study also analyzed mortality and factors associated with the mortality. A prediction model for mortality was constructed using a decision tree. Results: During the 4 fiscal years of the study period, from October 2016 - September 2020, a total of 287,449 individuals were diagnosed with acute appendicitis, and 272,850 appendectomies were performed, which gave an annual incidence (AI) of 10.8/10,000 person-years. The peak age incidence was in the second decade of life, which had an age-specific AI of 28.4/10,000 person-years. Using a multiplicative model of time series detrending and a decomposition model and a seasonal variation of acute appendicitis, with the highest incidence from May to October, linear regression showed a significant correlation between the incidence of appendicitis and precipitation (r2=0.28, p-value<0.001). The AI of appendicitis was declining and showed an especially large decline in the year 2020, when the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic began. The overall mortality was 2.3%. A significantly higher mortality rate was found in cases with extreme age (pediatric or geriatric), those with co-morbidities (diabetes, hypertension, dyslipidemia, chronic pulmonary disease, chronic kidney disease), and those with complications, especially generalized peritonitis, sepsis, and acute kidney injury. A decision-tree prediction model suggested that sepsis and renal complication were key nodes determining mortality risk in pediatric and geriatric appendicitis. Conclusion: A mortality risk from appendicitis remains in Thai patients, especially the elderly with co-morbidities and those with generalized peritonitis, sepsis, or renal complications. Early diagnosis and improvement in perioperative care in the extreme age groups might improve the mortality figure.

Lifestyle Influence on Mild Cognitive Impairment Progression: A Decision Tree Prediction Model Study.

This study assessed the influences of different lifestyle on mild cognitive impairment (MCI) progression and established a decision tree prediction model to analyse their predictive significance on MCI progression incidence. From October 2015 to February 2020,330 patients with MCI were recruited, and demographic and lifestyle information collected. They were followed up for 19.04 ± 10.227 months. Cognitive function was assessed using the Mini-Mental State Examination Scale every 6 months, and they were divided into MCI stable group and MCI progression group. The Kaplan Meier survival analysis showed an overall cohort survival rate of 33.2%; the annual conversion rate of MCI progression was 20%. Physical exercise, social engagement, high-fat diet, age, napping, and tea drinking were decision tree prediction model nodes. Hobbies were the most important factor for predicting MCI progression. The MCI progression probability rates were: with hobbies 26.829% (44 cases), without hobbies 57.831% (96 cases); for those withot hobbies, with physical exercise 43.077% (28 cases) without physical exercise 72.340% (68 cases); for those without hobbies with physical exercise and social engagement 20.000% (4 cases), without social engagement 53.333% (24 cases); for those without hobbies, physical exercises and social engagement and with nap habits 48.485% (16 cases), without nap habits 66.667% (8 cases). The decision tree prediction model AUC for predicting the MCI progression receiver operating characteristic curve was 0.737 (95% confidence interval: 0.685-0.785) (75.71% sensitivity, 71.75% specificity, P < 0.001. Hobbies, physical exercise, social engagement, napping, and drinking tea can help prevent MCI progression, while a high-fat diet may exacerbate MCI progression. In this study the rule with the lowest MCI progress probability for those who had hobbies, high-fat diet, and social engagement. And the decision tree model had good prediction efficiency.

Classification and Regression Tree Predictive Model for Acute Kidney Injury in Traumatic Brain Injury Patients.

Acute kidney injury (AKI) is prevalent in hospitalized patients with traumatic brain injury (TBI), and increases the risk of poor outcomes. We designed this study to develop a visual and convenient decision-tree-based model for predicting AKI in TBI patients. A total of 376 patients admitted to the emergency department of the West China Hospital for TBI between January 2015 and June 2019 were included. Demographic information, vital signs on admission, laboratory test results, radiological signs, surgical options, and medications were recorded as variables. AKI was confirmed since the second day after admission, based on the Kidney Disease Improving Global Outcomes criteria. We constructed two predictive models for AKI using least absolute shrinkage and selection operator (LASSO) regression and classification and regression tree (CART), respectively. Receiver operating characteristic (ROC) curves of these two predictive models were drawn, and the area under the ROC curve (AUC) was calculated to compare their predictive accuracy. The incidence of AKI on the second day after admission was 10.4% among patients with TBI. Lasso regression identified five potent predictive factors for AKI: glucose, serum creatinine, cystatin C, serum uric acid, and fresh frozen plasma transfusions. The CART analysis showed that glucose, serum uric acid, and cystatin C ranked among the top three in terms of the feature importance of the decision tree model. The AUC value of the decision-tree predictive model was 0.892, which was higher than the 0.854 of the LASSO regression model, although the difference was not statistically significant. The decision tree model is valuable for predicting AKI among patients with TBI. This tree-based flowchart is convenient for physicians to identify patients with TBI who are at high risk of AKI and prompts them to develop suitable therapeutic strategies.

Preoperative Delirium Risk Screening in Patients Undergoing a Cardiac Surgery: Results from the Prospective Observational FINDERI Study

ObjectivesPostoperative delirium (POD) is a common complication of cardiac surgery that is associated with higher morbidity, longer hospital stay, cognitive decline, and mortality. Preoperative assessments may help to identify patients´ POD risk. However, a standardized screening assessment for POD risk has not been established. DesignProspective observational FINd DElirium RIsk factors (FINDERI) study. ParticipantsPatients aged ≥ 50 years undergoing cardiac surgery. MeasurementsThe primary aim was to analyze the predictive value of the Delirium Risk Screening Questionnaire (DRSQ) prior to cardiac surgery. Secondary aims are to investigate cognitive, frailty, and geriatric assessments, and to use data-driven machine learning (ML) in predicting POD. Predictive properties were assessed using receiver operating characteristics analysis and multivariate approaches (regularized LASSO regression and decision trees). ResultsWe analyzed a data set of 504 patients (68.3 ± 8.2 years, 21.4% women) who underwent cardiac surgery. The incidence of POD was 21.0%. The preoperatively administered DRSQ showed an area under the curve (AUC) of 0.68 (95% CI 0.62, 0.73), and the predictive OR was 1.25 (95%CI 1.15, 1.35, p<0.001). Using a ML approach, a three-rule decision tree prediction model including DRSQ (score>7), Trail Making Test B (time>118), and Montreal Cognitive Assessment (score ≤ 22) was identified. The AUC of the three-rule decision tree on the training set was 0.69 (95% CI 0.63, 0.75) and 0.62 (95% CI 0.51, 0.73) on the validation set. ConclusionBoth the DRSQ and the three-rule decision tree might be helpful in predicting POD risk before cardiac surgery.

Risk prediction model of clinical mastitis in lactating dairy cows based on machine learning algorithms

Mastitis is the most common disease among dairy cows and is known to have negative effects on both animal welfare and the profitability of dairy farms. Early detection of clinical mastitis cases is considered the best option for preventing cows from developing mastitis. In this study, we developed clinical mastitis prediction models that only required inputting common indicators from the automatic milking system. We utilized multidimensional data from the cow mastitis database of Afimilk (China) Agricultural Technology Co., Ltd. to predict mastitis in dairy cows. All data were screened for the period of 0–150 days of lactation. The data included parity, lactation day, period, mean and standard deviation of milk yield, of electrical conductivity, and of lying time, which were taken as input features. The classification of whether cows suffer from clinical mastitis was determined as output. We analyzed 426 cows with clinical mastitis and 2087 healthy cows by using four machine learning algorithms: Decision Tree, Random Forest, Back Propagation neural networks, and Support Vector Machines. In these four algorithms, the accuracy ranged from 94% to 98%, while the running times varied widely from seconds to minutes. The decision tree prediction model achieved an accuracy of 98% and the precision rate for healthy cows was 99%, while for mastitis cows it was 97%. Machine learning algorithms have played an important role in predicting cow mastitis, with the Decision Tree algorithm showing great performance and higher accuracy in our research.

Analysis of the Influence of Parents Educational Level on Students Test Scores Based on Decision Tree

Under the background of the double reduction policy promulgated in July 2021 and the Peoples Republic of China Family Education Promotion Law promulgated in January 2022, improving family education and promoting home-school cooperation have become the focus of Chinas education reform and the key to improve the quality of students. The dataset of this paper comes from the Kaggle website. By using data visualization technology and ID3 algorithm of decision tree, the influence of various indicators on the accuracy of the decision tree prediction model is compared, and the relationship between parents education level and students test scores is further explored. The final results show that the education level of parents has a certain correlation with the test scores of children, and the students whose parents have higher education level generally have better scores.

Intelligent prediction model of mechanical properties of ultrathin niobium strips based on XGBoost ensemble learning algorithm

Ultrathin niobium strips with different thicknesses are prepared by an accumulative rolling process. The tensile test of the ultrathin niobium strips is carried out, and the microstructure of each niobium strip is characterized by electron backscattered diffraction (EBSD). The process parameters, mechanical properties and microstructure characterization data of rolled ultrathin niobium strips with different thicknesses are collected, analyzed and sorted. A data-driven intelligent prediction model for the mechanical properties of ultrathin niobium strips is established by deeply integrating the mechanical properties of ultrathin niobium strips with the microstructure evolution mechanism and combining the integrated data with the XGBoost ensemble learning algorithm. The optimal parameters of the XGBoost model are determined by a grid search and used for mechanical performance prediction. The overall generalization performance of the model is evaluated by the coefficient of determination (R2), mean absolute error (MAE), root mean square error (RMSE) and mean absolute percentage error (MAPE). To reflect the advancement of the proposed model, the prediction results of this model are compared with those of the random forest (RF), multi-layer perceptron (MLP) and gradient boosting decision tree (GBDT) prediction model. The research results show that, on the test set, the R2 in terms of the tensile strength and yield strength predicted values based on the XGBoost algorithm were higher than those of other models, reaching 0.944 and 0.964, respectively. The three error indicators corresponding to the XGBoost model were also at the lowest level. This means that the model based on the XGBoost algorithm has the optimal generalization performance and can thus realize the accurate prediction of the mechanical properties of ultrathin niobium strips. This research provides a new method and idea for the optimal design of rolling process and the prediction of their mechanical properties.

Mitochondrial mass of circulating NK cells as a novel biomarker in severe SARS-CoV-2 infection

BackgroundSevere SARS-CoV-2 infection results in lymphopenia and impaired function of T, B, and NK (TBNK-dominant) lymphocytes. Mitochondria are essential targets of SARS-CoV-2 and the efficacy of lymphocyte mitochondrial function for immunosurveillance in COVID-19 patients has not been evaluated. MethodsMulti-parametric flow cytometry was used to characterize mitochondrial function, including mitochondrial mass (MM) and low mitochondrial membrane potential (MMPlow), in TBNK-dominant lymphocytes from severe (n = 93) and moderate (n = 77) hospitalized COVID-19 patients. We compared the role of novel lymphocyte mitochondrial indicators and routine infection biomarkers as early predictors of severity and death in COVID-19 patients. We then developed a mortality decision tree prediction model based on immunosurveillance indicators through machine learning. ResultsAt admission, the MM of circulating NK cells (NK-MM) was the best discriminator of severe/moderate disease (AUC = 0.8067) compared with the routine infection biomarkers. The NK cell count and NK-MM displayed superior diagnostic effects to distinguish patients with non-fatal or fatal outcomes. Interestingly, NK-MM was significantly polarized in non-survivors, with some patients showing a decrease and others showing an abnormal increase. Kaplan-Meier analysis showed that NK-MM had the optimal predictive efficacy (hazard ratio = 11.66). The decision tree model has the highest proportion of importance for NK-MM, which is superior to the single diagnostic effect of the above indicators (AUC = 0.8900). ConclusionNK-MM was not only associated with disease severity, its abnormal increases or decreases also predicted mortality risk. The resulting decision tree prediction model is the first to focus on immune monitoring indicators to provide decision-making clues for COVID-19 clinical management.

Prediction of the hand function part of the Fugl‐Meyer scale after stroke using an automatic quantitative assessment system

AbstractHand function assessment is an essential component of the process of stroke rehabilitation because of the high incidence of hand motor dysfunction. In terms of the manual evaluation of hand function, the Fugl‐Meyer scale is a recommended scale with high reliability and validity. However, the need for accurate assessments and increasing developments in technology has led to the promotion of automatic quantitative assessment systems for the hand. In this study, we collected quantitative data on hand function with an automatic system and the upper limb Fugl‐Meyer assessment (FMA) from 79 people with stroke. We developed decision tree (DT) and gradient‐boosted decision tree (GBDT) predictive models for the Fugl‐Meyer score using features extracted from the Hand Automatic Quantitative Assessment System (HAQAS). Predictive performances were compared between these models regarding the predictive accuracy and Cohen's kappa. There were high correlations between features automatically collected by the HAQAS and the Fugl‐Meyer scale in all the sub‐items, with the maximal correlations all being over 0.5, indicating the high validity of the HAQAS in automatic FMA prediction. Hand functions were more highly correlated (average correlation coefficient 0.90) with HAQAS features than wrist functions (average correlation coefficient 0.54), and the GBDT achieved higher predictive accuracies and agreement than the DT algorithm. We conclude that the HAQAS is feasible for stroke patients with hand dysfunction and convenient for clinicians and therapists. This study was registered in the Chinese Clinical Trial Registry (ChiCTR1800019098).

Decision tree predictive model for dimensional control of side flange bearing housings

Decision tree predictive model for dimensional control of side flange bearing housings

Establishment of Decision Tree Prediction Model for Risk Factors of Placenta Accreta Spectrum Disorders

To analyze the risk factors for placenta accreta spectrum (PAS) disorders and to construct preliminarily a decision tree prediction model for PAS, to help identify high-risk populations, and to provide reference for clinical prevention and treatment. By accessing the electronic medical record system, we retrospectively analyzed the relevant data of 2022 women who gave birth between January 2020 and September 2020 in a hospital in Chengdu. Univariate logistic regression and multivariate logistic regression were conducted to analyze the risk factors of PAS. SPSS Clementine12.0 was used to make preliminary exploration for the decision tree prediction model of PAS risk factors. Results of logistic regression suggested that the top three risk factors for PAS included the following, the risk of PAS in pregnant women with placenta previa was 8.00 times that in pregnant women without placenta previa (95% CI: 5.24-12.22), the risk of PAS in multiple pregnancies was 2.52 times that in singleton pregnancies (95% CI: 1.72-3.69), and the risk of PAS in pregnant women who have had three or more abortions was 1.89 times that in those who have not had abortion (95% CI: 1.11-3.20). Results of the decision tree prediction model based on C5.0 algorithm were as follows, placenta previa was the most important risk factor, with as high as 93.33% (140/150) patients developed PAS when they had placenta previa; when in vitro fertilization-embryo transfer (IVF-ET) was the only factor the subjects had, the incidence of PAS was 59.91% (133/222); the incidence of PAS was as high as 75.96% (79/104) when the subjects had both IVF-ET and a history of uterine surgery; the probability of PAS in women who had induced abortion in the past was 48.46% (205/423); the probability of PAS in women who had undergone uterine surgery previously was 10.54% (37/351); the incidence of PAS was as high as 100.00% (163/163) when the subjects had induced abortion previously and uterine surgery history. The model showed a prediction accuracy of 85.41% for the training set and a prediction accuracy of 83.36% for the testing set, both being high rates of accuracy. The decision tree prediction model can be used for rapid and easy screening of patients at high risk for PAS, so that the likelihood of PAS can be actively and dynamically assessed and individualized preventive measures can be taken to avoid adverse outcomes.

Portfolio Optimization Techniques for Cryptocurrencies

This article addresses the shortcomings of the existing literature regarding cryptocurrency portfolio construction. First, we address the effectiveness of time-series models that capture stylized features. We perform a comparison study on various methods for estimating distributions for asset returns, including normal, historical, and GARCH models within a CVaR setting. The goal of this comparison is to determine the financial benefits of constructing portfolios based on estimated distributions that consider stylized features of crypto return series. Next, we create and compare various prediction models for cryptocurrencies and integrate them with mean-variance optimization to base performance on portfolio management metrics, such as Sharpe ratio and level of diversification, rather than statistical metrics like accuracy and R² on which the literature solely focuses. We determine it is unclear which optimization approach (CVaR or Robust MVO) leads to better crypto portfolios, and so, to address this, we compare optimization procedures on out-of-sample data through a thorough cross-validation of hyperparameters for each technique. We then compare the resulting risk-optimal portfolios from each technique. The results show that a CVaR approach with a GARCH simulation and a decision tree prediction model with robust mean-variance optimization yield portfolios of similar risk. We also show that using statistical metrics to evaluate models may not always yield the best financial performance.

Decision Tree Prediction Model in Patient Mortality Rate based on Risk Factors

The Coronavirus disease (Covid-19) has become a global problem since WHO declared a pandemic in 2020. The number of deaths due to Covid-19 has increased significantly in many countries. This study aimed to implement decision tree modeling to represent the relationship between risk factors and the mortality rate of Covid-19 patients. This study analyzed secondary data of 83,024 Covid patients from January 2020 to June 2021. Data processing used data mining with the decision tree classification method. The results showed that comorbidity is the leading risk factor for death which is then influenced by age. The higher the age group with comorbidities, the higher the risk of death. Suggested that health services can utilize the results of this study to prevent the severity of Covid-19 infection. Such as the development of comorbid awareness programs and community-based education on managing patients with comorbidities.

A diagnostic model for predicting type 2 nasal polyps using biomarkers in nasal secretion.

Predicting type 2 chronic rhinosinusitis with nasal polyps (CRSwNP) may help for selection of appropriate surgical procedures or pharmacotherapies in advance. However, an accurate non-invasive method for diagnosis of type 2 CRSwNP is presently unavailable. To optimize the technique for collecting nasal secretion (NasSec), 89 CRSwNP patients were tested using nasal packs made with four types of materials. Further, Th2low and Th2highCRSwNP defined by clustering analysis in another 142 CRSwNP patients using tissue biomarkers, in the meanwhile, inflammatory biomarkers were detected in NasSec of the same patients collected by the selected nasal pack. A diagnostic model was established by machine learning algorithms to predict Th2highCRSwNP using NasSecs biomarkers. Considering the area under receiver operating characteristic curve (AUC) for IL-5 in NasSec, nasal pack in polyvinyl alcohol (PVA) was superior to other materials for NasSec collection. When Th2low and Th2highCRSwNP clusters were defined, logistic regression and decision tree model for prediction of Th2highCRSwNP demonstrated high AUCs values of 0.92 and 0.90 respectively using biomarkers of NasSecs. Consequently, the pre-pruned decision tree model; based on the levels of IL-5 in NasSec (≤ 15.04 pg/mL), blood eosinophil count (≤ 0.475*109/L) and absence of comorbid asthma; was chosen to define Th2lowCRSwNP from Th2highCRSwNP for routine clinical use. Taken together, a decision tree model based on a combination of NasSec biomarkers and clinical features can accurately define type 2 CRSwNP patients and therefore may be of benefit to patients in receiving appropriate therapies in daily clinical practice.

Ontology-Based decision tree model for prediction of fatty liver diseases

Non-Alcohol Fatty liver disease is a common clinical complication. The paper aimed to develop a knowledge-based fatty liver detection system based on an ontology and detection rules extracted from a decision tree algorithm. Ontology is created to represent knowledge related to patients and fatty liver disease. By utilizing 43 SWRL rules and the Drool inference engine in ontology, we detected fatty liver patients. The training dataset size is 70% of clean data, including 580 electronic medical records of patients who suffer from liver diseases. After inferencing the rules, the number of patients who suffer from fatty liver disease in ontology is the same as the decision tree model. The paper validated the result generated by the ontology model through the results of the decision tree model.

A prospective study of the decision tree prediction model for attention deficit hyperactivity disorder in preschool children.

To study the clinical value of attention time combined with behavior scale in the screening of attention deficit hyperactivity disorder (ADHD) in preschool children. A total of 200 preschool children with ADHD diagnosed in Fujian Maternal and Child Health Hospital from February 2019 to March 2020 were enrolled as the ADHD group. A total of 200 children who underwent physical examination in the hospital or kindergartens during the same period were enrolled as the control group. Attention time was recorded. Chinese Version of Swanson Nolan and Pelham, Version IV Scale-Parent Form (SNAP-IV) scale was used to evaluate symptoms. With clinical diagnosis as the gold standard, the decision tree analysis was used to evaluate the clinical value of attention time combined with behavior scale in the screening of ADHD. Compared with the control group, the ADHD group had significantly higher scores of SNAP-IV items 1, 4, 7, 8, 10, 11, 14, 15, 16, 18, 20, 21, and 22 (P<0.05) and a significantly shorter attention time (P<0.05). The variables with statistically significant differences between the two groups in univariate analysis were used as independent variables to establish a decision tree model. The accuracy of the model in predicting ADHD was 81%, that in predicting non-ADHD was 69%, and the overall accuracy was 75%, with an area under the ROC curve of 0.816 (95% CI: 0.774-0.857, P<0.001). The decision tree model for screening ADHD in preschool children based on attention time and assessment results of behavior scale has a high accuracy and can be used for rapid screening of ADHD among children in clinical practice.

Predicting and Preventing Crime: A Crime Prediction Model Using San Francisco Crime Data by Classification Techniques

The crime is difficult to predict; it is random and possibly can occur anywhere at any time, which is a challenging issue for any society. The study proposes a crime prediction model by analyzing and comparing three known prediction classification algorithms: Naive Bayes, Random Forest, and Gradient Boosting Decision Tree. The model analyzes the top ten crimes to make predictions about different categories, which account for 97% of the incidents. These two significant crime classes, that is, violent and nonviolent, are created by merging multiple smaller classes of crimes. Exploratory data analysis (EDA) is performed to identify the patterns and understand the trends of crimes using a crime dataset. The accuracies of Naive Bayes, Random Forest, and Gradient Boosting Decision Tree techniques are 65.82%, 63.43%, and 98.5%, respectively, and the proposed model is further evaluated for precision and recall matrices. The results show that the Gradient Boosting Decision Tree prediction model is better than the other two techniques for predicting crime, based on historical data from a city. The analysis and prediction model can help the security agencies utilize the resources efficiently, anticipate the crime at a specific time, and serve society well.

Predicting adoption of colorectal cancer screening among Korean Americans using a decision tree model

ABSTRACT Background Colorectal cancer screening (CRCS) rates remain suboptimal among Korean Americans despite recommendations from health organizations. Little is known about the mechanism underlying their CRCS adoption within complex systems. This study aimed to examine the multi-level predictors of CRCS adoption among Korean Americans using a decision tree model. Methods A cross-sectional survey was performed to assess CRCS adoption and multiple levels of influence – individual (i.e. CRCS self-efficacy, CRCS attitudes, risk of colorectal cancer, psychological distress, health status), interpersonal (i.e. social support, social networks, CRCS recommendations), and organizational/community (i.e. health insurance, primary doctor, primary clinic) factors. A total of 433 Korean Americans aged 50–75 in a metropolitan area in the Southeastern U.S. completed a self-report questionnaire. To determine the important variables that predict CRCS adoption, the study generated a decision tree predictive model using R statistical software. Results The results indicated that CRCS self-efficacy and CRCS attitudes at the individual level and CRCS recommendations and social support at the interpersonal level differentiate adopting or not adopting CRCS. Furthermore, CRCS recommendations (n = 138, 56%, prob = 0.64) and CRCS self-efficacy (n = 51, 21%, prob = 0.88) were the most powerful predictors of CRCS adoption. Conclusion The findings highlight the critical roles of CRCS recommendations from healthcare providers and family/friends and patients’ confidence in performing screening-related tasks in influencing CRCS adoption among Korean Americans. Practice efforts should target individual and interpersonal characteristics when developing interventions for promoting CRCS among Korean Americans, especially who are not adherent to screening guidelines.

Decision tree model predicts live birth after surgery for moderate-to-severe intrauterine adhesions

BackgroundAfter treatment of intrauterine adhesions, the rate of re-adhesion is high and the pregnancy outcome unpredictable and unsatisfactory. This study established and verified a decision tree predictive model of live birth in patients after surgery for moderate-to-severe intrauterine adhesions (IUAs).MethodsA retrospective observational study initially comprised 394 patients with moderate-to-severe IUAs diagnosed via hysteroscopy. The patients underwent hysteroscopic adhesiolysis from January 2013 to January 2017, in a university-affiliated hospital. Follow-ups to determine the rate of live birth were conducted by telephone for at least the first postoperative year. A classification and regression tree algorithm was applied to establish a decision tree model of live birth after surgery.ResultsWithin the final population of 374 patients, the total live birth rate after treatment was 29.7%. The accuracy of the model was 83.8%, and the area under the receiver operating characteristic curve (AUC) was 0.870 (95% CI 7.699–0.989). The root node variable was postoperative menstrual pattern. The predictive accuracy of the multivariate logistic regression model was 70.3%, and the AUC was 0.835 (95% CI 0.667–0.962).ConclusionsThe decision tree predictive model is useful for predicting live birth after surgery for IUAs; postoperative menstrual pattern is a key factor in the model. This model will help clinicians make appropriate clinical decisions during patient consultations.

Novel SVD integrated with GBDT based Virtual Sample Generation and Its Application in Soft Sensor

With the coming of the big data era, data-driven based modeling approaches have become the hot research topic in recent years. Unfortunately, due to the limitation of the actual process, the data is basically in a steady state and it is difficult to obtain enough high-quality data, which is defined as the small sample (SS) problem. Recently, to deal with the SS problem, a virtual sample generation (VSG) approach based on the distribution of the original data has been taken into account. In this paper, a VSG method based on singular value decomposition (SVD) feature decomposition and gradient boosting decision tree (GBDT) prediction model (SVD-GBDT) is proposed. In the proposed SVD-GBDT method, firstly, the distribution characteristics of the original data are used to extract the main features and expand the number of samples by using the SVD algorithm. Then the GBDT algorithm is used to find the virtual output of the virtual samples by the SVD method. Finally, SVD and GBDT are combined to complete the sample expansion (SVD-GBDT-VSG). In this paper, we choose the purified terephthalic acid (PTA) industrial process to verify the effectiveness of the proposed methodology. Simulation results show that compared with related methods, the proposed SVD-GBDT-VSG algorithm in this paper can achieve sample expansion well and at the same time can effectively improve the accuracy performance of soft measurement.