In the semiconductor manufacturing industry, the need for continuous quality improvement has never been more pronounced. This demand is driven by an unprecedented influx of manufacturing data, with more than 1000 process parameters recorded for a single wafer, and tens of thousands of wafers being produced daily. Traditional statistical methods have proven insufficient to fully exploit these massive volumes of data. As such, this paper explores the application of hybrid machine learning techniques, specifically Memory Based Reasoning (MBR) and Neural Network (NN) learning, as more powerful tools for managing this complexity and improving yield in semiconductor manufacturing.