US5835891A — Device modeling using non-parametric statistical …

Claim 1

Independent Claim 181 words

A method for modeling an electronic component comprising the following steps: (a) measuring performance of the electronic component to produce a plurality of data vectors representing actual performance of the electronic component; (b) selecting boundary vectors from the plurality of data vectors, including using a computing system to perform the following substeps: (b.1) calculating a density estimate for each data vector in the plurality of data vectors, (b.2) partitioning a first selected number of the data vectors from the plurality of data vectors into a partitioned group, the partitioned group including the first selected number of data vectors from the plurality of data vectors with greatest density estimates, as determined in substep (b.1), (b.3) selecting a second selected number of the partitioned group as candidates for boundary vectors, the candidates for boundary vectors including the second selected number of the partitioned group with lowest density estimates, as determined in substep (b.1), and (b.4) selecting the boundary vectors from the candidates for boundary vectors; and, (c) from the boundary vectors, calculating values for boundary model parameter vectors which model the electronic component.

Claim 2

Independent Claim 181 words

A method for modeling an electronic component comprising the following steps: (a) measuring performance of the electronic component to produce a plurality of data vectors representing actual performance of the electronic component; (b) selecting boundary vectors from the plurality of data vectors, including using a computing system to perform the following substeps: (b.1) calculating a density estimate for each data vector in the plurality of data vectors, (b.2) partitioning a first selected number of the data vectors from the plurality of data vectors into a partitioned group, the partitioned group including the first selected number of data vectors from the plurality of data vectors with greatest density estimates, as determined in substep (b.1), (b.3) selecting a second selected number of the partitioned group as candidates for boundary vectors, the candidates for boundary vectors including the second selected number of the partitioned group with lowest density estimates, as determined in substep (b.1), and (b.4) selecting the boundary vectors from the candidates for boundary vectors; and, (c) from the boundary vectors, calculating values for boundary model parameter vectors which model the electronic component.

Claim 3

Dependent Claim 66 words

A method as in claim 1 additionally comprising the following steps: (d) from the density estimate for each data vector calculated in substep (b.1), selecting a nominal data vector as a data vector from the plurality of data vectors which has a highest density estimate; and, (e) from the nominal data vector, calculating a value for a nominal model parameter vector which models the electronic component.

Claim 4

Dependent Claim 122 words

A method as in claim 1 wherein substep (b.1) comprises performing the following substeps for each data vector in the plurality of data vectors, for the performance of the following substeps, each data vector being a current data vector: (b.1.1) locating a third selected number of data vectors from the plurality of data vectors which are nearest to the current data vector; (b.1.2) calculating an average distance from the current data vector to each of the third selected number of data vectors from the plurality of data vectors which are nearest to the current data vector found in substep (b.1.1); and, (b.1.3) estimating a density estimate for the current data vector as one divided by the average distance calculated in substep (b.1.2).

Claim 5

Dependent Claim 31 words

A method as in claim 3 wherein in substep (b.1.1) the third selected number is equal to a total number of data vectors in the plurality of data vectors minus one.

Claim 6

Dependent Claim 44 words

A method as in claim 1 wherein step (b) additionally comprises the following substep performed before substep (b.1): scaling vector components of all data vectors within the plurality of data vectors so that the plurality of data vectors are bounded by a unit hypercube.

Claim 7

Dependent Claim 22 words

A method as in claim 1 wherein substep (b.4) includes selecting all the candidates for the boundary vectors as the boundary vectors.

Claim 8

Dependent Claim 104 words

A method as in claim 1 wherein substep (b.4) comprises the following substeps: (b.4.1) for all possible pairings of the candidates for boundary vectors, selecting a pairing of candidates for boundary vectors which are closest together; (b.4.2) removing, from the candidates, a candidate from the pairing of candidates selected in substep (b.4.1) which has a higher density estimate as calculated in substep (b.1); (b.4.3) for candidates remaining after performance of substep (b.4.2), repeating substeps (b.4.1) and (b.4.2) until a number of remaining candidates is equal to a third selected number; and, (b.4.4) selecting the third selected number of remaining candidates as the boundary vectors.

Claim 9

Independent Claim 192 words

A method for modeling an electronic component comprising the following steps: (a) measuring performance of the electronic component to produce a plurality of data vectors representing actual performance of the electronic component; (b) from the plurality of data vectors, calculating values for a plurality of model parameter vectors which model the electronic component; and, (c) selecting boundary vectors from the plurality of model parameter vectors, including using a computing system to perform the following substeps: (c.1) calculating a density estimate for each model parameter vector in the plurality of model parameter vectors, (c.2) partitioning a first selected number of the model parameter vectors from the plurality of model parameter vectors into a partitioned group, the partitioned group including the first selected number of model parameter vectors from the plurality of model parameter vectors with greatest density estimates, as determined in substep (c.1), (c.3) selecting a second selected number of the partitioned group as candidates for boundary vectors, the candidates for boundary vectors including the second selected number of the partitioned group with lowest density estimates, as determined in substep (c.1), and (c.4) selecting the boundary vectors from the candidates for boundary vectors.

Claim 10

Independent Claim 192 words

A method for modeling an electronic component comprising the following steps: (a) measuring performance of the electronic component to produce a plurality of data vectors representing actual performance of the electronic component; (b) from the plurality of data vectors, calculating values for a plurality of model parameter vectors which model the electronic component; and, (c) selecting boundary vectors from the plurality of model parameter vectors, including using a computing system to perform the following substeps: (c.1) calculating a density estimate for each model parameter vector in the plurality of model parameter vectors, (c.2) partitioning a first selected number of the model parameter vectors from the plurality of model parameter vectors into a partitioned group, the partitioned group including the first selected number of model parameter vectors from the plurality of model parameter vectors with greatest density estimates, as determined in substep (c.1), (c.3) selecting a second selected number of the partitioned group as candidates for boundary vectors, the candidates for boundary vectors including the second selected number of the partitioned group with lowest density estimates, as determined in substep (c.1), and (c.4) selecting the boundary vectors from the candidates for boundary vectors.

Claim 11

Dependent Claim 49 words

A method as in claim 8 additionally comprising the following steps: (d) from the density estimate for each model parameter vector calculated in substep (c.1), selecting a nominal model parameter vector as a model parameter vector from the plurality of model parameter vectors which has a highest density estimate.

Claim 12

Dependent Claim 134 words

A method as in claim 8 wherein substep (c.1) comprises performing the following substeps for each model parameter vector in the plurality of model parameter vectors, for the performance of the following substeps, each model parameter vector being a current model parameter vector: (c.1.1) locating a third selected number of model parameter vectors from the plurality of model parameter vectors which are nearest to the current model parameter vector; (c.1.2) calculating an average distance from the current model parameter vector to each of the third selected number of model parameter vectors from the plurality of model parameter vectors which are nearest to the current model parameter vector found in substep (c.1.1); and, (c.1.3) estimating a density estimate for the current model parameter vector as one divided by the average distance calculated in substep (c.1.2).

Claim 13

Dependent Claim 33 words

A method as in claim 10 wherein in substep (c.1.1) the third selected number is equal to a total number of model parameter vectors in the plurality of model parameter vectors minus one.

Claim 14

Dependent Claim 47 words

A method as in claim 8 wherein step (c) additionally comprises the following substep performed before substep (c.1): scaling vector components of all model parameter vectors within the plurality of model parameter vectors so that the plurality of model parameter vectors are bounded by a unit hypercube.

Claim 15

Dependent Claim 104 words

A method as in claim 8 wherein substep (c.4) comprises the following substeps: (c.4.1) for all possible pairings of the candidates for boundary vectors, selecting a pairing of candidates for boundary vectors which are closest together; (c.4.2) removing, from the candidates, a candidate from the pairing of candidates selected in substep (c.4.1) which has a higher density estimate as calculated in substep (c.1); (c.4.3) for candidates remaining after performance of substep (c.4.2), repeating substeps (c.4.1) and (c.4.2) until a number of remaining candidates is equal to a third selected number; and, (c.4.4) selecting the third selected number of remaining candidates as the boundary vectors.

Claim 16

Dependent Claim 157 words

Computer Readable Storage media which stores software for generating boundary vectors from a data set consisting of a plurality of data vectors, the software when executed on a computing system comprising: calculating means for calculating a density estimate for each data vector in the plurality of data vectors; partitioning means for partitioning a first selected number of the data vectors from the plurality of data vectors into a partitioned group, the partitioned group including the first selected number of data vectors from the plurality of data vectors with greatest density estimates, as determined by the calculating means; number selecting means for selecting a second selected number of the partitioned group as candidates for the boundary vectors, the candidates for the boundary vectors including the second selected number of the partitioned group with lowest density estimates, as determined by the calculating means; and, vector selecting means for selecting the boundary vectors from the candidates for the boundary vectors.

Claim 17

Dependent Claim 157 words

Computer Readable Storage media which stores software for generating boundary vectors from a data set consisting of a plurality of data vectors, the software when executed on a computing system comprising: calculating means for calculating a density estimate for each data vector in the plurality of data vectors; partitioning means for partitioning a first selected number of the data vectors from the plurality of data vectors into a partitioned group, the partitioned group including the first selected number of data vectors from the plurality of data vectors with greatest density estimates, as determined by the calculating means; number selecting means for selecting a second selected number of the partitioned group as candidates for the boundary vectors, the candidates for the boundary vectors including the second selected number of the partitioned group with lowest density estimates, as determined by the calculating means; and, vector selecting means for selecting the boundary vectors from the candidates for the boundary vectors.

Claim 18

Dependent Claim 99 words

Storage media as in claim 14 wherein, for each data vector in the plurality of data vectors in turn being a current vector, the calculating means: locates a third selected number of data vectors from the plurality of data vectors which are nearest to the current data vector, calculates an average distance from the current data vector to each of the third selected number of data vectors from the plurality of data vectors which are nearest to the current data vector, and estimates a density estimate for the current data vector as one divided by the calculated average distance.

Claim 19

Dependent Claim 28 words

Storage media as in claim 15 wherein the third selected number is equal to a total number of data vectors in the plurality of data vectors minus one.

Claim 20

Dependent Claim 14 words

Storage media as in claim 14 wherein the calculating means utilizes a K-D tree.

Claim 21

Dependent Claim 68 words

Storage media as in claim 14, wherein the software, when executed on a computing system, additionally comprises: scaling means for scaling vector components of all data vectors within the plurality of data vectors so that the plurality of data vectors are bounded by a unit hypercube, the scaling being performed before the calculating means calculates the density estimate for each data vector in the plurality of data vectors.

Claim 22

Dependent Claim 23 words

Storage media as in claim 14 wherein the vector selecting means selects all the candidates for the boundary vectors as the boundary vectors.

Claim 23

Dependent Claim 108 words

Storage media as in claim 14 wherein the vector selecting means includes: selecting means for selecting, for all possible pairings of the candidates for boundary vectors, a pairing of candidates for boundary vectors which are closest together; and, removing means for removing, from the candidates, a candidate from the pairing of selected candidates which has a higher density estimate as calculated by the calculating means; wherein the selecting means selects pairs of candidates and the removing means removes candidates until a number of remaining candidates is equal to a third selected number, the vector selecting means selecting the third selected number of remaining candidates as the boundary vectors.

Annuity Stage	Window Opens	Payment Deadline	Surcharge Period Until	Estimated Fee	Status
3.5 Years (Year 4)	2001-11-09	2002-05-11	2002-11-10	$2000.00	Window Closed / Lapsed
7.5 Years (Year 8)	2005-11-09	2006-05-11	2006-11-10	$3760.00	Window Closed / Lapsed
11.5 Years (Year 12)	2009-11-09	2010-05-11	2010-11-10	$7700.00	Window Closed / Lapsed

Device modeling using non-parametric statistical determination of boundary data vectors

Abstract

Inventors

Bibliographic Details

Claims Analysis

References Cited (Backward Citations)

USPTO Annuity & Maintenance Windows

Assigned Categories & Tags

Assign Category Tag