The Discriminant Analysis for the Distribution of Wafer Defect
碩士 === 銘傳大學 === 應用統計資訊學系碩士班 === 101 === The spatial random defects in the wafer are usually caused by the particle cluster in the atmosphere or due to some process that does not meet the required specification. Thus, there should be some spatial pattern for the defects in the wafer. Since the spatia...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2013
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/22148622896112856468 |
| Summary: | 碩士 === 銘傳大學 === 應用統計資訊學系碩士班 === 101 === The spatial random defects in the wafer are usually caused by the particle cluster in the atmosphere or due to some process that does not meet the required specification.
Thus, there should be some spatial pattern for the defects in the wafer. Since the spatial pattern could result imprecise estimation of the wafer yield, the recognition of the clusters of the spatial patterns are important. In this paper, we use the auto-logistic regression model (ALRM) and cubic splines function to fit the distribution of wafer defect. We also use the maximum pseudo-likelihood estimation (MPLE) distinguish the different pattern of Wafer Defect. Finally, the real data analysis is used to verify our model and comparison with classification and regression trees (CART) algorithm.
|
|---|