The objectives of this research are to propose a proper model for defective products in an autoparts manufacturing industry to apply the proposed model to the defective products data to investigate the factors related to the machines producing defective products and to compare the proposed model with the generalized estimating equations model (GEE). The proposed model is a generalized linear mixed model (GLMM) in which the dependent variables have a Poisson distribution. The parameter are estimated using a Bayesian method. The data were collected from 12 machines 15 days for each machine from an autoparts manufacturing factory. The data collection dates vary depending on the working steps and the product types. The results show that the proposed model fits the data. The factors related to the machines producing defective products are working steps workers and product types. According to the estimates of the factors effects the working step producing the largest number of defective products is the working step 2 followed by the working steps 3 1 and 4 respectively. The worker producing the largest of defective products is the worker 4 followed by the workers 10 5 8 6 2 3 1 7 and 9 respectively. The product type 3 has the largest number of defective products followed by the product types 2 and 1 respectively. If the population-averaged effects are focused the GEE is applied. However the subject-specific is focused the GLMM is applied.

This research proposes a hidden Markov model for detecting machines with multiple fillers producing soft drink for quality control in order to reduce defective products. The proposed model is compared with a classical generalized estimating equations model (GEE). The monthly data the number of defective products the correctness and the time spending in correction were collected from a machine in a sample factory in Pratumthanee province. The two hidden states of the fillers are workable and unworkable. The states has a logistic distribution. The results show that the correctness of the filler error occurrence does not improve the probability of being in a unworkable state neither the period of time spending on correction. The current state seems to be the same as the previous state

This research proposes a Bayesian logistic regression model which is applied to the data from autoparts manufacturing machines. Factors related to defective and bad products are investigated. The proposed model is compared with the logistic regression using maximum likelihood method for parameter estimation. The data were collected from 132 machines in an autoparts manufacturing factory. The research found that useful life machine type 6 worker group 3 and 4 working step 1 and 2 influence to the risk of producing defective and bad products. When the useful life is increased by 1 month the risk of producing defective and bad products will be increased by 2.2%. The risk that the machine type 6 will produce defective and bad products is 4.078 times greater than the risk that the machine type will do. The risk that the worker group 3 will produce defective and bad products is 61.7% less than the risk that the worker group 12 will do. The risk that the worker group 4 will produce defective and bad products is 61.5% less than the risk that the worker group 12 will do. The risk that the working step 1 will produce defective and bad products is 2.831 times greater than the risk that the working step 4 will do. The risk that the working step 2 will produce defective and bad products is 13.8 % greater than the risk that the working step 4 will do. The parameter estimates from the Bayesian logistic regression are very close to the ones from the logistic regression using maximum likelihood method for parameter estimation