An Application of Bootstrapping in Logistic Regression Model

Adjei, Isaac Akpor; Karim, Rezaul

doi:10.4236/oalib.1103049

Cited by 16 publications

(16 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The medical categories were clustered into 5 larger categories based on their prevalence in the current sample: mental health, musculoskeletal and injury, neurological, cardiovascular and blood disease, and other . These categories were entered into the analysis as predictors, and regression coefficients were estimated based on a bootstrapping procedure with 5000 successful replicates [ 17 ]. Logistic regression analyses were performed using the R-based statistical software JASP (JASP version 0.13.1).…”

Section: Methodsmentioning

confidence: 99%

Perceived Impact of Outdoor Swimming on Health: Web-Based Survey

Massey¹,

Gorczynski²,

Harper³

et al. 2022

Interact J Med Res

View full text Add to dashboard Cite

Background Outdoor swimming in lakes, lidos (outdoor pools), rivers, and the sea has grown in popularity in many countries, including the United Kingdom. Many anecdotal accounts indicate improvements in medical conditions, which are considered a consequence of outdoor swimming. Objective The aim of this study is to better understand outdoor swimmers’ perceptions of their health and the extent to which participation impacted their existing self-reported symptoms. Methods A survey was conducted to investigate outdoor swimming behaviors and reports of any diagnosed medical conditions. Medical conditions were coded into categories, and descriptive statistics were generated regarding the outdoor swimmers’ behaviors and the effect that outdoor swimming had on their medical symptoms if any. The medical categories were clustered into five larger categories based on their prevalence in the current sample: mental health; musculoskeletal and injury; neurological; cardiovascular and blood disease; and other, which comprises inflammatory, immune, endocrine, and respiratory conditions. Results In total, 722 outdoor swimmers responded, of whom 498 (68.9%) were female. The probability of outdoor swimming having some positive impact on health across all medical categories was 3.57 times higher compared with no impact (B=1.28, 95% CI 0.63-1.91; P<.001), 44.32 times higher for the mental health category (B=3.79, 95% CI 2.28-5.30; P<.001), 5.25 times higher for musculoskeletal and injury category (B=1.66, 95% CI 0.52-2.79; P=.004), and 4.02 times higher for the other category (B=1.39, 95% CI 0.27-2.51; P=.02). Overall, outdoor swimming was associated with perceived reductions in symptoms of poor mental health (χ22=25.1; P<.001), musculoskeletal and injury (χ22=8.2; P=.04), cardiovascular and blood (χ22=14.7; P=.006), and other conditions (χ22=18.2; P<.001). Conclusions Physical activity in the form of outdoor swimming is perceived to have positive impacts on health and is associated with perceived symptom reductions in mental health, musculoskeletal and injury, and cardiovascular and blood conditions. This study cannot provide causal relationships or provide mechanistic insights. However, it does provide a starting point for more targeted prospective intervention research into individual conditions or categories of conditions to establish the impact in those who choose to start outdoor swimming.

show abstract

Section: Methodsmentioning

confidence: 99%

Perceived Impact of Outdoor Swimming on Health: Web-Based Survey

Massey¹,

Gorczynski²,

Harper³

et al. 2022

Interact J Med Res

View full text Add to dashboard Cite

show abstract

“…Research on the tendency of the HVB area to show more HBV in rural areas than urban areas [15], [16]. In analyzing HVB risk factors, one of the analysis models that can be used is logistic regression [12], [17].…”

Section: Discussionmentioning

confidence: 99%

“…In general, the odds ratio (odds ratios) is a set of opportunities divided by other opportunities. The odds ratio for predictors is defined as the relative number where the probability of yield increases (odds ratio> 1) or down (odds ratio <1) when the predictor variable value increases by 1 unit [12].…”

Section: Methodsmentioning

confidence: 99%

Analysis of the Relationship between Traditional Birth Attendant Service and the Risk of Hepatitis B Transmission among Pregnant Women at the Indonesian Rural Community

Bustan¹,

Ansariadi²,

Syam³

et al. 2019

Proceedings of the Proceedings of the 3rd International Conference on Environmental Risks and Public Health, ICER-PH 2018, 26-2

View full text Add to dashboard Cite

Childbirth assistance with traditional birth attendant is carried out without regard to the sterilization factor so that transmission of the hepatitis virus can occur. This study aimed to determine the likelihood of Hepatitis B transmission in pregnant women based on the history of delivery assistance by traditional birth attendants. The study design was a case-control study of 120 cases of positive HVBsAg pregnant women and 120 negative HVBsAg pregnant women controls from the HVB screening program of Health Centers throughout Bulukumba District, Indonesia. Binary logistic regression analysis was carried out to determine the magnitude of the relationship between risk factors for labor history by TBA and HVB infection. Bivariate logistic regression model found that pregnant women with a history of childbirth by TBA have a risk of transmitting HVB 1.8 times greater (Odds Ratio = 1.78; 95% CI 1.07-2.98) compared with those who do not have a history of labor for TBA. The multivariate logistic regression model showed that pregnant women who received TBA help who had a history of blood transfusion and a history of surgical treatment had a risk of contracting HVB by 1.8 times greater than those delivered by midwives or other skilled health workers (Odds Ratio = 1.882; 95% CI 1,078-3,080). TBA services that pay less attention to sterilization in childbirth services have a risk of HVB transmission, especially if it occurs in those who have a history of blood transfusion and surgery.

show abstract

“…Several methods have been prescribed in the literature, including, (a) the separation of the dataset into training subset to first build the model, and testing subsets to validate it; (b) k-fold cross-validation (CV) involving splitting the dataset into k-number of (roughly) equally-sized subsets for model development and validation; (c) a computationally more-expensive version of the k-fold CV, known as the leave-one-out-cross-validation or LOOCV; and (d) bootstrapping with replacement [59][60][61][62]. The suitability and limitations of these approaches have been well studied, albeit with no general consensus or recommendations (e.g., [63][64][65][66]). In this study, we have adopted a variant to the k-fold CV approach for model testing by randomly splitting the dataset into training (95%) and testing (5%) 500 times.…”

Section: Logistic Regression (Lr) Model Development and Evaluationmentioning

confidence: 99%

Reliable Predictors of Arsenic Occurrence in the Southern Gulf Coast Aquifer of Texas

Venkataraman

Lozano

2018

Geosciences

View full text Add to dashboard Cite

Arsenic contamination of groundwater in the Southern Gulf Coast Aquifer of Texas is a critical public health concern as much of the area is rural in nature with decentralized water supplies. Previous studies have pointed to volcanic deposits as the regional source of arsenic but no definitive or reliable predictors of arsenic maximum contaminant level (MCL) exceedance have been identified. In this study, we have studied the effect of various hydrogeochemical parameters as well as soil and land-use variables on arsenic MCL exceedance using logistic regression (LR) techniques. The LR models display good accuracy of 75% or higher but suffer from a high rate of false negatives, highlighting the challenges in capturing the spatial irregularities of arsenic in this region. Despite not displaying high statistical significance, pH appears to be an important variable in the LR models-its effect on arsenic exceedance is not clear and warrants further investigation. The results of the study also show that groundwater vanadium and fluoride are consistently the only significant variables in the models developed; the positive coefficients for both these elements indicates a common geogenic source for arsenic, fluoride and vanadium, corroborating the findings of earlier studies.

show abstract

An Application of Bootstrapping in Logistic Regression Model

Cited by 16 publications

References 11 publications

Perceived Impact of Outdoor Swimming on Health: Web-Based Survey

Perceived Impact of Outdoor Swimming on Health: Web-Based Survey

Analysis of the Relationship between Traditional Birth Attendant Service and the Risk of Hepatitis B Transmission among Pregnant Women at the Indonesian Rural Community

Reliable Predictors of Arsenic Occurrence in the Southern Gulf Coast Aquifer of Texas

Contact Info

Product

Resources

About