|
Objective To explore the community-based family planning, public health agencies to carry out the feasibility of AIDS VCT services. BP and GA-BP neural network in MATLAB software, and compare the two network fitting data effect to determine the network model, using a sensitivity analysis of the factors affecting the technical staff of family planning, health departments jointly launched the AIDS VCT service attitude, community IPPF , Health jointly carry out the feasibility of VCT services to provide a theoretical basis for the prevention and control of AIDS for the government to provide a basis for decision making. Research Methods field studies using a combination of qualitative and quantitative way. Topics interviews and group interviews of 60 family planning, health management and technical personnel to understand the grassroots to HIV / AIDS / VCT cognitive and attitude, analyzed birth control, health departments jointly launched VCT services enabling factors that may be encountered obstacles and countermeasures. Using cluster sampling, the three communities of 437 meter of Health, the Health Service staff face-to-face asking using structured questionnaire to collect information using χ2 test, variance analysis, non-parametric test analysis community technology officer AIDS knowledge, voluntary consultation detection (VCT) knowledge situation. Fitting this paper data prepared based on MATLAB BP, the GA-BP neural network program, MSE, EPOCH, R2, adjusted R2 for network performance evaluation the comparison network performance to determine the network model. By MATLAB based sensitivity analysis, come to affect the technical staff of the IPPF, the health sector is carried out jointly by the degree of VCT service attitude factors. Research results at the management level, the family planning network and improve, \\\Factors favor at the technical level, the Community IPPF, health departments jointly launched VCT services is the most vulnerable community work with VCT, \The lack of knowledge of the breadth, depth, staff mobility, lack of staff, inadequate hardware facilities and other factors that may be barriers to community family planning, public health departments jointly launched VCT services. Interviewees believe birth control, the health sector in issuing VCT promotional materials, HIV pre-test counseling, condom promotion, HIV antibody screening \SRH related consulting \More than 70% of the interviewees believe that in VCT HIV testing, half of the interviewees do not know VCT free, less than 1/3 of the object know free HIV testing can be done at all levels of CDC. More than 60% of people believe that \80% of the interviewed missionaries participated in AIDS voluntary counseling and testing services, qualitative interview results show that the VCT community service impact factors are mainly four aspects: the social environment (discrimination in society, and society can be understood accepted), the results are confidential, service and, quality of service. Total score of 18.48 ± 3.39 the 437 respondents AIDS knowledge VCT knowledge sharing razors total score of the technical staff of about 10.09 ± 1.76. 30% do not know or do not know \courtesy kissing \Community technician anal sex with HIV () syringe sharing, HIV () \less than 5%, to accept HIV () blood products infected with HIV, the probability of awareness more than 70%. Knowledge awareness of the technical staff of the \Χ2 test shows that when the hidden layer neurons for 5, 10, 15, 20 when the BP the GABP neural networks fit the data reach the number of pre-set target of 0.01 to reach the number of pre-set target iteration step 1000, statistics have Differences between (P lt; 0.05). With the increase in the number of hidden layer neurons, BP, GA-BP neural network to fit the data to achieve the target number of iteration steps 1000 steps to reach the number of pre-set target to increase (P lt; 0.001). Wilcoxon rank sum test: when the hidden layer neurons for 5, 10, 15, 20, GA-BP neural network adjustment R2, R2 are greater than BP neural network. The Kruskal-Wallis test results show that: the number of hidden layer neurons is 15, the coefficient of determination R2 of the BP neural network and GA-BP neural network have reached the maximum. Of hidden layer neurons number 15, GA-BP neural network to fit the data. GA-BP neural network to fit the data results, the community grassroots technicians titles, the average monthly income, AIDS knowledge score, the VCT knowledge score, work department, unit level, working hours, education level IPPF Health carried out jointly by the VCT services greater impact (sensitivity values ??were greater than 0.60). Conclusion Community of Health and Health jointly carry out the feasibility of VCT services the grassroots to staff VCT knowledge needed further training in order to work as a community VCT backbone. The titles of the primary technical staff, the average monthly income, AIDS knowledge score, score VCT knowledge, work department, unit level, working hours, education level affect their IPPF, health carried out jointly by the VCT service attitude. BP neural network genetic algorithm to optimize the initial weights and threshold values, to overcome the slow convergence of BP neural network into local minima problems can better fit the data; neural networks, sensitivity analysis can render the impact of the input variables.
|