| Summary: | 碩士 === 國立成功大學 === 工業與資訊管理學系專班 === 95 === The European Union issued Directive on the restriction and use of certain hazardous substances in electrical and electronic equipment. Known as Restriction of Hazardous Substances (RoHS), it was enforced throughout the European community effective 1 July 2006. With the RoHS directive, the companies whose electronic products that include lead based soldering would not be allowed to trade with the member states of the European Union. Therefore, companies would begin to experiment on a reengineering effort that could be unsuccessful if they do not find suitable alternatives.
The data is collected from the surface mount process station in IC assembly factory while the company phased in lead-free process. Model tree is used to construct the quality control model to help process engineers identifying the critical variables that affect the results as yield. To validate the model, back-propagation neural network and multiple linear regression are used in this research.
Finally, the predictive capability of this model is similar with back-propagation neural network but much better than multiple linear regression. Besides, it indeed provides clear process information and predicted yield for engineers through the tree branches and regression functions in the leaf-nodes while making decisions on troubleshooting and experiment designs.
|
|---|
