隨著社會的不斷進步,消費者對於產品或服務品質的要求標準也越來越高。 無疑的,品質已成為消費者選擇物品或服務時所需考慮的重要條件之一,同時也是決定企業能否永續經營發展的關鍵點。品質已成為企業保有競爭優勢的重要利基,製程品質管制更是影響產品品質的關鍵。品質的提升,品質管理、品質改善與統計手法三者必需連為一體,而且有效地相互配合運用,才能發掘潛在的問題點。一般運用在品質管制上的統計方法有假設檢定、抽樣檢驗、實驗計劃法、相關迴歸、品管七大手法…等。 本文主要在探討列陣(Array)製程中的A檢雷射repair機台修補基板成功率與相關重要因素的關係。希望藉由統計分析了解可能會影響修補成功率的變因狀況,進而分析影響修補成功率的重要因子,並將此重要因子提供工程做製程改善以提高修補良率及提高良品率,進而減少損失及減少成本以提升品質的競爭力。 資料分析探討的部分是針對TFT雷射修補成功率與修補條件背景變數做基本資料的描述;其次以重要因素做分析,首先使用卡方獨立性檢定,探討兩變數之間的相關性高低,再針對變數做關聯性分析;最後使用邏輯斯回歸模型作模型的選擇,並決定其最終模型以及製作預測機率圖並加以解釋。
As society progresses, the consumer product or service quality standards are also getting higher and higher. Undoubtedly, quality has become a consumer choice of goods or services that need to be considered one of the important elements, but also deciding whether the development of sustainable management of key points. Quality has become an important competitive advantage to maintain niche, process quality control is the key to the impact of product quality. Enhance the quality, quality management, quality improvement and statistical methods necessary for the three together, and the effective use of each other in order to identify potential problems. Generally used in statistical quality control methods on the Hypothesis Testing, Sampling Inspection, Design of Experiment, Regression, Quality 7 Tools ... etc. This article discusses Array laser repair process. The relationship between the success ratio of the Array laser repair equipment and the related important factors. To realize the factor that may affect the repair success ratio, and analyze the important factors of the success ratio through statistical analysis. This research provides this important factor to do the system improvement and enhance yield. To reduce loses and cost, to improve the quality. The material analysis discussion's part is aims at the TFT laser repair success ratio and the repair condition background variable. And analyze the important factors. First, It discuss two varies relevance by using chi-squared independence test. Secondly, it do relational analysis to varies. Finally, to choose model by Logistic Regression Model. It decides and explains the final model and Estimated Probability Plot.