Advanced Methods for the Analysis of Semiconductor Manufacturing Process Data - Advanced Techniques in Knowledge Discovery and Data Mining - page 67

Table 2.5. Physical and technological meaning of selected variables.

Parameter number Explanation

IO device
78, 79 Threshold voltages

Logic NMOS device
114, 118, 120, 129 Threshold voltages

115, 119, 121 Saturation currents

131 Punchthrough current

Logic PMOS device
140, 143 Threshold voltages

141, 144 Saturation currents

142 Channel leakage

Parameters unrelated to split
1 Breakdown voltage

32 Saturation current MV device

65 Sheet resistance well

71, 73 Threshold voltage HV devices

191 Gate oxide thickness

198, 199, 201, 205 Sheet resistance poly

Derived parameter
126 Universal curve FOM

that NOVCLASS only uses the samples a liated to class 1 of the training set
during learning. Thus, samples a liated to classes 2 and 3 were not involved
in the training of NOVCLASS and were unknown to the OCC classifier. In
the following, classes 2 and 3 will be merged to class 2, denoting abnormal or
novel measurements and respective process states. The aim was to assess the
feasibility of NOVCLASS for (semi)automatic significance and novelty data

Fig. 2.24. Visualization of selected SPLIT6 classification.