Expert knowledge methods and systems for data analysis

A method for desensitizing a process variable associated with a wafer processing operation, the desensitizing configured to prevent the process variable from causing a false positive fault that can cause the wafer processing operation to halt, comprising:

• referencing an original model representative of the processing operation, the original model including variable data and normally distributed data of the processing operation;
• running a new processing operation with a new variable to generate new variable data representative of the new processing operation;
• projecting the new variable data onto the original model;
• examining data points identified to be outliers as a result of the projecting, an outlier being indicative of a fault that should cause processing to be halted;
• receiving expert knowledge to ascertain whether the process variable that caused the data point to be an outlier is unimportant;
• generating desensitizing data using a randomizer to normalize the new variable data using normally distributed random data, the variable data of the original model, and the new variable data;
• and
• augmenting data that was used to generate the original model with the desensitizing data, the augmenting configured to prevent the process variable from causing the data point to be falsely identified as an outlier in a subsequent processing operation, wherein each operation of the method is executed by a processor.