Artificial Neural Network Modeling Of Female Stress Urinary Incontinence In Southern Xinjiang

Object: Through to our country the influence factors of female stress urinary incontinence on Meta analysis of risk factors as the neural network model of the preset parameters,combined with the southern xinjiang region of female stress urinary incontinence relevant data forecast model is set up,in order to prevent and reduce the occurrence of female stress urinary incontinence and provides reference to improve the quality of life of women.Methods: Eight databases including Pub Med,Web of Science,The Cochrane Library,Clinical Key,Wanfang Data,CNKI,VIP,and Sinomed were used to retrieve literature studies on the risk factors related to stress urinary incontinence in Chinese women.Retrieval time limits are established from the database until May 2020.In strict accordance with the inclusion and exclusion criteria,two researchers independently screened and extracted the literature data.After the risk of bias assessment of the included literature,Rev Man 5.3software was used for Meta-analysis,so as to obtain the risk factors of stress urinary incontinence in adult women in China.Through the southern xinjiang region pelvic floor functional disorder diagnosis and treatment of the new technology promotion and the related research on clinical application of integrated research on China’s southern xinjiang region of adult female pelvic floor functional disorder epidemiological investigation and obtain the relevant data about the part stress urinary incontinence.Epidata 3.1 software was used for to double the original data entry and check each other,SPSS 25.0 was used to filtered raw data and analysis,IBM SPSS Modeler Subscription was used to multilayer perceptron neural network prediction model is set up,at the same time,the model is tested and validated.Results: A total of 35 studies,including 94043 subjects,were included in the Meta-analysis,involving 20 risk factors.The results showed: Age [OR =2.39,95%CI(2.02,2.82),P<0.00001],labor intensity [OR =1.48,95%CI(1.30,1.69),P <0.00001],Body mass index(BMI)[OR =1.61,95%CI(1.44,1.81),P <0.00001],The history of drinking [OR=1.38,95%CI(1.25,1.53),P <0.00001],hypertension [OR =1.68,95%CI(1.35,2.08),P<0.00001],constipation [OR =1.62,95%CI(1.46,1.80),P <0.00001],history of respiratory system [OR =2.46,95%CI(2.15,2.81),P <0.00001],personal history of diseases of the genitourinary system [OR =2.31,95%CI(1.90,2.80),P <0.00001],history of gynecological diseases [OR =3.02,95%CI(1.35,6.72),P =0.007],pregnancy [OR =1.33,95%CI(1.08,1.63),P <0.006],times of births(≥ 3 times)[OR =1.53,95%CI(1.39,1.68),P <0.00001],weight of the first fetus [OR =2.10,95%CI(1.13,3.91),P =0.02],perineal laceration [OR =1.51,95%CI(1.32,1.73),P <0.00001],vaginal delivery [OR =1.91,95%CI(1.55,2.36),P <0.00001],pausimenia [OR =1.80,95%CI(1.48,2.19),P <0.00001],history of pelvic surgery [OR =2.15,95%CI(1.47,3.15),P <0.0001],uterine prolapse [OR =2.21,95%CI(1.68,2.91),P <0.00001]are all risk factors for stress urinary incontinence in Chinese women,and cesarean section[OR =0.71,95%CI(0.51,0.97),P =0.03] are Protective factors for stress urinary incontinence in Chinese women.A total of 928 samples were collected in southern Xinjiang,of which 368 were excluded from males and missing data such as height and weight.560 samples were eventually included for the model construction,among which 216 were diagnosed with stress urinary incontinence.A multi-layer perceptron(MLP)neural network model was established by using SPSS Modeler Subsubscription.560 cases of samples were randomly divided into training,testing and validation groups according to 70%,15%,and 15%(repeatable random allocation was selected,and the random seed was 7462243.The input predictive variables were Meta-analysis results,and the output target was stress urinary incontinence.Model building options: Build a standard model with two hidden layers,the first have 20 neurons and the second have 15 neurons.The termination condition is one.Accuracy reaches 90%.2.or the model could not continue to reduce the error,so the reproducible results were selected,and the random seed was 477534151.The results showed that the accuracy of the model reached 86.8%,and the order of importance of the predictive variables indicated age,BMI,the weight of the first fetus,uterine prolapse,pregnancy number,perineal laceration,etc.The model was verified by 75 samples,and the results showed a sensitivity was 93.1% and a specificity was 80.43%,the accuracy was 85.33%,the AUC was 0.924,and the Gini coefficient was 0.848.Conclusions: 1.Current evidence suggests that the age,body mass index(BMI),the intensity of labor,history of drinking,hypertension,constipation,respiratory disease history,personal history of diseases of the genitourinary system,history of disease of department of gynecology,pregnant time,production time(≥3),the first child fetal weight,perineal laceration,vaginal birth,menopause,history of pelvic surgery,uterine prolapse is a risk factor for female stress urinary incontinence in China,and cesarean section is the protection of the Chinese female stress urinary incontinence.Chinese adult women and medical workers should pay attention to the above risk factors,and individualized intervention or treatment should be carried out for patients with different risk factors,in order to reduce the incidence of stress urinary incontinence in Chinese women and improve the quality of life of such patients.2.The artificial neural network model of female stress urinary incontinence in southern Xinjiang has a good prediction effect,which can provide reference value for individuals and physicians in the prevention,treatment and other interventions of stress urinary incontinence.