Using fuzzy logistic regression for modeling vague status situations: Application to a dietary pattern study

Distribution mixtures are used as models to analyze grouped data. The estimation of parameters is an important step for mixture distributions. The latent class model is generally used as the analysis of mixture distributions for discrete data. In this paper, we consider the parameter estimation for a mixture of logistic regression models. We know that the expectation maximization (EM) algorithm was most used for estimating the parameters of logistic regression mixture models. In this paper, we propose a new type of fuzzy class model and then derive an algorithm for the parameter estimation of a fuzzy class logistic regression model. The effects of the explanatory variables on the response variables are described. The focus is on binary responses for the logistic regression mixture analysis with a fuzzy class model. An algorithm, called a fuzzy classification maximum likelihood (FCML), is then created. The mean squared error (MSE) based accuracy criterion for the FCML and EM algorithms to the parameter estimation of logistic regression mixture models are compared using the samples drawn from logistic regression mixtures of two classes. Numerical results show that the proposed FCML algorithm presents good accuracy and is recommended as a new tool for the parameter estimation of the logistic regression mixture models.

A novel technique for parameter estimation in intuitionistic fuzzy logistic regression model

Fuzzy QSARs for predicting log Koc of persistent organic pollutants

Fuzzy logistic regression based on the least squares approach with application in clinical studies

To evaluate the contrast-enhanced ultrasound(CEUS) and ultrasonic elastography (UE) in diagnosis of thyroid nodules by a binary Logistic regression model. A total 149 cases with thyroid nodules were examed by CEUS and UE and were confirmed by surgical pathology. A Logistic model was obtained on the basis of ultrasonographic features. Receiver operator characteris(ROC) curve was constructed to assess the performance of the Logistic model. Four ultrasonographic features including calcification, enhancement degree, biggest perfusion strength and elastic score were finally entered into the Logistic model. The percentage of correct prediction was 91.90%. Both CEUS and UE have high value in differential diagnosis of thyroid nodules and the Logistic regression model has high diagnostic rate.

Background/Objectives: Different methods can be applied to create predictive models for the clinical data with binary outcome variable. This research aims to explore the process of constructing the modified predictive model of Logistic Regression (LR). Method/Statistical Analysis: To improve the accuracy of prediction, the Distance based Outlier Detection (DBOD)is used for pre-processing and Bipolar Sigmoid Function calculated using Neuro based Weight Activation Function is used in Logistic Regression instead of Sigmoid Function. Datasets were collected from clinical laboratory of AR Hospital in Madurai for the three years 2012, 2013 and 2014 are used for analysis. Data pre-processing is done to avoid the existence of insignificant data in the dataset. The detected outliers, using DBOD method are treated using a method closest to the normal range. A comparative study among different distance measures likes Euclidean and Manhattan etc. are done for DBOD method. The pre-processed data finally is fed as input to the Logistic Regression model. Maximum likelihood estimation is used to fit the model. Logistic Model is built from the Sigmoid Function using the Regression Coefficients. The accuracy of the model is evaluated by 10 fold cross validation. Findings: Logistic Model is built from the Sigmoid Function using the Regression Coefficients, produces the accuracy of 79%. The Sigmoid Function calculated using Random Weight Function provides the prediction accuracy of 84.2% and the Bipolar Sigmoid Function calculated using Neuro based Weight Activation function provides the prediction accuracy of 90.4%. On comparison, Bipolar Sigmoid Function calculated using Neuro weight activation function outperforms well than the Sigmoid Function calculated using regression coefficients. Improvements/Applications: The accuracy of Logistic Regression is improved from 79% to 90.4%. The most important factors: Erythrocyte Sedimentation Rate (ESR) and Estimation of Mean blood Glucose are identified from positive subjects of Diabetes Mellitus. The analysis is done for the 31 Diabetes Disease attributes of three years dataset. Keywords: Bipolar Sigmoid Neuro-Weight Activation Function, Distance based Outlier Detection Method, Logistic Regression, Random Weight Function, Sigmoid Activation Function, Type 2 Diabetes Risk Factors

Objective: This study aims to identify data-driven dietary patterns and find out the relationship of dietary patterns with different cardiometabolic risks including overweight/obesity, dyslipidemia or high blood pressure (HBP) among older adult males (aged ≥65 years old) residents (>10 years) of Sichuan, Southwestern China. Methods: This cross-sectional study investigated the relationships of dietary patterns with cardiometabolic risks in older adult males (n=1136) aged ≥65 years old referred to the Public Hospital in Sichuan Province, China. A validated semi-quantitative food frequency questionnaire (SFFQ) was used to investigate the correlations between the animal-based and processed food dietary (pattern 1), balanced dietary (pattern 2), and Ovo-Lacto Vegetarian dietary (pattern 3) and different cardiometabolic risks. The correlations between dietary patterns and the laboratory tests were calculated. Results: The animal-based and processed food, balanced, and Ovo-Lacto Vegetarian dietary patterns accounted for 14.83%, 14.36% and 11.86% of the variance, respectively. The Chi-Square test showed the dietary patterns were associated with triglycerides (TG) (P < 0.001), total cholesterol (TC), Systolic blood pressure (SBP), and body mass index (BMI). Adjusted logistic regression analysis showed the animal-based and processed food dietary pattern was positively associated with overweight/obesity (OR: 3.25, 95% CI: 1.94, 5.46) and dyslipidemia (OR: 3.53, 95% CI: 2.00, 6.22). The balanced dietary pattern was negatively associated with overweight/obesity (OR: 0.51, 95% CI: 0.36, 0.72), dyslipidemia (OR: 0.50, 95% CI: 0.35, 0.75) and HBP (OR: 0.54, 95% CI: 0.38, 0.77). The Ovo-Lacto Vegetarian dietary pattern was negatively associated with dyslipidemia (OR: 0.56, 95% CI: 0.39, 0.81) and hyperuricemia (OR: 0.56, 95% CI: 0.39, 0.79). No statistically significant relationship was found between the three dietary patterns and impaired fasting glucose (P: 0.05). Conclusions: Elderly males in Sichuan Province with overweight/obesity, dyslipidemia and high blood pressure seem to have adopted different dietary patterns.

Dietary and social behaviour are non-medical factors that influence health outcomes. Non-communicable diseases are related to dietary patterns. To date, little is known about how social behaviour is associated with health-related dietary patterns, and, in particular, we lack information about the role of sex within this possible relation. Our cross-sectional study investigated associations between dietary patterns and social behaviour including personality traits (self-control, risk taking), political preferences (conservative, liberal, ecological, social) and altruism (willingness to donate, club membership, time discounting) in men and women. We performed sex-specific correlation analyses to investigate relationships between dietary patterns based on self-reported protocols from the Mediterranean Diet Adherence Screener (MEDAS) and the validated Healthy Eating Index (HEI) from the EPIC Study and a self-reported social behaviour questionnaire. In linear regression models, we analysed associations between dietary and social behaviour patterns. Sex differences were measured by interaction analysis for each social behaviour item. The study sample consisted of N = 102 low-risk individuals. The median age of the study participants was 62.4 (25th/75th percentile 53.6, 69.1) years, and 26.5% were women. Analyses showed that a lower HEI score was correlated with a higher BMI in both women and men. MEDAS and HEI showed a positive correlation with each other in men. In men, a higher MEDAS showed a positive correlation when they estimated their ability as high, with the same for self-control and preference for ecological politics and MEDAS. A weak negative correlation has been shown between men with a preference for conservative politics and MEDAS. HEI showed a positive significant correlation with age in men. Male participants without club membership scored significantly higher in the HEI compared to non-members. A negative correlation was shown for time discounting in men. Linear regression models showed positive associations between preferences for ecological-oriented politics and nutrition for both HEI and MEDAS. No sex interactions were observed. We faced a few limitations, such as a small sample size, particularly for women, and a limited age spectrum in a European cohort. However, assuming that individuals with a preference for ecological-oriented politics act ecologically responsibly, our findings indicate that ecological behaviour in low-risk individuals might determine, at least in part, a healthy diet. Furthermore, we observed dietary patterns such as higher alcohol consumption in men or higher intake of butter, margarine and cream in women that indicate that women and men may have different needs for nutritional improvement. Thus, further investigations are needed to better understand how social behaviour affects nutrition, which could help to improve health. Our findings have the potential to inform researchers and practitioners who investigate the nature of the relationship between social behaviour and dietary patterns to implement strategies to create first-stage changes in health behaviour for individuals with a low cardiovascular risk profile.

Exercise tests are considered the most reliable ones for the preoperative workup of lung resection candidates but frequently are indicated only in cases with low predicted postoperative forced expiratory volume in 1s (FEV1) and carbon monoxide diffusing capacity (DLCO). The aim of this investigation is to evaluate if a simple, standardised incremental bicycle exercise test routinely performed in patients considered operable is predictive of postoperative cardio-respiratory complication and if its performance is comparable to a logistic regression model including frequently cited clinical predictive variables. A series of 103 lung resection candidates were included in a prospective observational study. All patients underwent a standardised, incremental exercise test on a bicycle up to exhaustion. The analysed outcome was the occurrence of postoperative cardio-respiratory complications prospectively recorded and codified. The correlation of distance reached at the end of the test and the outcome was estimated by non-parametric tests. A logistic regression model including age, BMI, predicted postoperative (ppoFEV1%) and predicted postoperative DLCO (ppoDLCO) was adjusted and the individual probability of complication calculated and set as a new variable. Finally, two receiver operating characteristic (ROC) curves were constructed and compared: one with distance at the exercise test and the other with logistic regression probability of complication. Mortality of the series was nil. Outcome prevalence was 14%. Distance reached at the end of the exercise test was lower in cases with a positive outcome (3498.6m vs 4543.5m, p=0.001). On logistic regression analysis, age of the patient (p=0.016) and ppoDLCO (p=0.000) were predictive for the outcome. On ROC analysis, C-indices were 0.77 (distance) and 0.78 (logistic model, p=0.95). Reached distance in a simple standardised exercise test is related to postoperative morbidity after lung resection; and the accuracy of prediction using this variable alone is comparable to a logistic regression model including age and ppoDLCO.

Bispectral analysis (BIS) of the electroencephalogram (EEG) has been shown in retrospective studies to predict whether patients will move in response to skin incision. This prospective multicenter study was designed to evaluate the real-time utility of BIS in predicting movement response to skin incision using a variety of general anesthetic techniques. Three hundred patients from seven study sites received an anesthetic regimen expected to give an approximately 50% movement response at skin incision. EEG was continuously recorded via an Aspect B-500 monitor and BIS was calculated in real time from bilateral frontocentral channels displayed on the monitor. Half of the patients were randomized to a treatment group in which anesthetic drug doses were increased to produce a lower BIS. In the control group, BIS was recorded, but no action taken on the data displayed. A determination of movement in response to skin incision was made in the 2 min succeeding incision. Retrospective pharmacodynamic modeling was performed using STANPUMP to estimate effect-site concentrations of intravenously administered anesthetics. BIS values were significantly higher in the control group (66 +/- 19) versus the BIS-guided group, in which additional anesthesia was administered to produce a lower BIS (51 +/- 19). The movement response rate was significantly higher in the control group at 43% compared with 13% in the BIS-guided group, but response rates were low at sites which used larger doses of opioids. Logistic regression analysis showed that BIS, estimated opioid effect-site concentrations, and heart rate (in that order) were the best predictors of movement at skin incision. This study demonstrates that dosing anesthetic drugs to lower BIS values achieves a lower probability of movement in response to surgical stimulation. BIS is a significant predictor of patient response to incision, but the utility of the BIS depends on the anesthetic technique being used. When drugs such as propofol or isoflurane are used as the primary anesthetic, changes in BIS correlate with the probability of response to skin incision. When opioid analgesics are used, the correlation to patient movement becomes much less significant, so that patients with apparently "light" EEG profiles may not move or otherwise respond to incision. Therefore, the adjunctive use of opioid analgesics confounds the use of BIS as a measure of anesthetic adequacy when movement response to skin incision is used as the primary end point.

Statistical Prediction in Pathological Types of Chronic Kidney Disease.

Background and aim: Prostate cancer is the most common cancer among men. Several studies have investigated the effects of dietary patterns on prostate cancer risk, but this topic is still a matter of debate. This study aimed to examine the association between dietary patterns and prostate cancer risk.Methods: In a case-control study, 60 newly diagnosed prostate cancer cases and 60 hospital-based controls were selected from two main hospitals of Shiraz, Iran. Data on dietary intakes, anthropometric features, and demographic characteristics were collected. To determine the dietary patterns factor analysis, and to estimate the odds ratios (ORs), multivariable logistic regression was performed.Results: Two major dietary patterns were identified: Western dietary (WD) pattern and Mediterranean dietary (MD) pattern. After adjusting for potential confounders, men who had higher scores for WD pattern (above the median) were more likely to have prostate cancer (OR = 5.15; 95% CI (1.44–18.47); P = 0.01) compared with men who had lower scores. A nonsignificant inverse association was found for MD pattern (OR = 0.62; 95% CI (0.22–1.77); P = 0.37).Conclusions: Our findings suggest that WD pattern may increase the risk of prostate cancer and the beneficial effects of MD pattern on prostate cancer risk need further research.

Summary form only given. This paper focuses on the potential demand of a product and proposes three models adopting following each case: (1) a model in which the potential market size does not change with time (conventional logistic model); (2) a model in which the potential market size increases with time stepwise (n-step logistic model); and (3) a model in which the potential market size increases with time continuously (double logistic model). The paper confirms the validity of the models by comparing the predicted values obtained by nonlinear regression analysis with the cases of three consumer electronic products.

Differential item functioning (DIF) is a pernicious statistical issue that can mask true group differences on a target latent construct. A considerable amount of research has focused on evaluating methods for testing DIF, such as using likelihood ratio tests in item response theory (IRT). Most of this research has focused on the asymptotic properties of DIF testing, in part because many latent variable methods require large samples to obtain stable parameter estimates. Much less research has evaluated these methods in small sample sizes despite the fact that many social and behavioral scientists frequently encounter small samples in practice. In this article, we examine the extent to which model complexity—the number of model parameters estimated simultaneously—affects the recovery of DIF in small samples. We compare three models that vary in complexity: logistic regression with sum scores, the 1-parameter logistic IRT model, and the 2-parameter logistic IRT model. We expected that logistic regression with sum scores and the 1-parameter logistic IRT model would more accurately estimate DIF because these models yielded more stable estimates despite being misspecified. Indeed, a simulation study and empirical example of adolescent substance use show that, even when data are generated from / assumed to be a 2-parameter logistic IRT, using parsimonious models in small samples leads to more powerful tests of DIF while adequately controlling for Type I error. We also provide evidence for minimum sample sizes needed to detect DIF, and we evaluate whether applying corrections for multiple testing is advisable. Finally, we provide recommendations for applied researchers who conduct DIF analyses in small samples.

Background and Objectives: Traumatic duodenal injury is a rare disease with limited evidence. We aimed to evaluate the risk factors for postoperative leakage and outcomes of pyloric exclusion after duodenal grade 2 and 3 injury. Materials and Methods: We reviewed a prospectively collected trauma database for the period January 2004–December 2020. Patients with grade 2 and 3 traumatic duodenal injury were included. To identify the risk factors for postoperative leakage, we used a stepwise multivariable logistic regression model and a least absolute shrinkage and selection operator (LASSO) logistic model. We constructed a receiver operator characteristic (ROC) curve to predict risk factors for postoperative leakage. Results: During the 17-year period, 179,887 trauma patients were admitted to a regional trauma center in Korea. Of these patients, 74 (0.04%) had duodenal injuries. A total of 49 consecutive patients had grade 2 and 3 traumatic duodenal injuries and underwent laparotomy. The incidence of postoperative leakage was 32.6% (16/49). Overall mortality was 18.4% (9/49). A stepwise multivariable logistic regression and LASSO logistic regression model showed that time from injury to initial operation was the sole statistically significant risk factor. The ROC curve at the optimal threshold of 15.77 h showed the following: area under ROC curve, 0.782; sensitivity, 68.8%; specificity, 87.9%; positive predictive value, 73.3%; and negative predictive value, 85.3%. There was no significant difference in outcomes between primary repair alone and pyloric exclusion. Conclusions: Time from injury to initial operation may be the sole significant risk factor for postoperative duodenal leakage. Pyloric exclusion may not be able to prevent postoperative leakage.