Yield Improvement Using Advanced Data Analytics

We are living in an era in which data is growing in an exponential pace and coming from multiple sources. This type of data has been called "Big Data". Big data has large volume, variety of formats, high dimensionality and the need for a high speed processing. Those features differentiates it from traditional datasets. Hence data management, analysis, visualization and results communications are getting more complex. The potential of obtaining greater knowledge and more actionable conclusions makes it very attractive. Therefore a data-driven mindset is emerging in different industries and the semiconductor industry is not an exception. This paper describes the results for yield improvement of our silicon carbide technology using advanced data analytics. In doing so, the paper outlines how the data was collected, managed and preprocessed to make it suitable for analysis. It explains which methods and algorithms were used to explore the data, uncover patterns and identify the most important features/predictors. At the end, challenges and conclusions are presented.