Advanced manufacturing insight system for semiconductor application

1. A method of reducing multitudes of input data signals to a manageable plurality of input data signals using a computer system and user feedback, comprising the steps of:

inputting each of the multitudes of input data signals to the computer system;

determining which of the multitudes of input data signals are heuristically significant input data signals using the computer system and a grading system that includes heuristical and/or statistical information;

further reducing the multitudes of input data signals to the manageable plurality of input data signals using the computer system and the heuristically significant input data signals via the computer system monitoring passive and active user feedback; and

using the manageable plurality of input data signals to obtain response data, wherein the input data signals are a subset of machine-readable entry sourced from a semiconductor wafer, packaging, or design facility, and wherein the response data is associated with the semiconductor wafer, the packaging, or the design facility.

2. The method according to claim 1 wherein the grading system includes weights for each of the following with respect to certain ones of the input signals:

strength of a correlation;

consistent slope across each of a plurality of tools, where each of the plurality of tools are electronic equipment electrically coupled to the computer system;

consistent slope across time; correlation of multiple variables with respect to the certain ones of the input signals; and

any other derived indicator of the input signal.

3. The method according to claim 2 wherein each of the weights are combined to provide the grading system.

4. The method according to claim 1 wherein determining actively significant input signals further includes differentiating user feedback between levels of positive and negative feedback, and asymmetrically assigning weights such that positive feedback is more heavily weighted than negative feedback.

5. The method according to claim 4 wherein the feedback from certain users are filtered.

6. The method according to claim 5 wherein a plurality of feedback occurrences that occur within a predetermined period of time are reduced to fewer occurrences.

7. The method according to claim 6 wherein determining actively significant input signals further includes volatility differentiation of user feedback between recently occurring user feedback and selectively aged user feedback.

8. The method according to claim 7 wherein the volatility and feedback differentiation is linearly weighted.

9. The method according to claim 7 wherein the volatility and feedback differentiation is non-linearly weighted.

10. The method according to claim 1 wherein the input data is summarized equipment fault detection and classification data and the response is yield.

11. The method according to claim 1 wherein the input data is summarized equipment fault detection and classification data from any process step(s) in either wafer or packaging areas, and the response is test chip parametric or yield data.

12. The method according to claim 1 wherein the input data and response data is a subset of a manufacturing execution system, in-line metrology, consumable, environmental, maintenance log or schedule, process characterization, wafer acceptance test, scribe, product bin, product parametric test, product yield, packaging, or non-lot data.