Data mining in electronics packaging

In this paper the most common methods of data mining have been investigated and their application in electronics will be reflected. The current developments are systematized by the type of the used techniques. Therefore, a comprehensive literature review of data mining in electronics has been accomplished. The paper describes the usage of data mining in defect cause analysis, effects of process parameter for quality, deployment of equipment and maintenance. Examples of data mining applications in the literature have been summarized. A comprehensive experimental setup was the basis for the investigation on the effects on void generation. Statistical analysis and data mining techniques were used to identify the main causes for voids. The data file encompasses materials, suppliers, process parameters and inspection results. For a detailed analysis the x-ray inspection data of voids has been clustered into groups according to the dedicated package type. Finally, a neural network approach is applied to the experimental data and the model results are discussed.