Ex Parte Fan et al

Patent Trial and Appeal Board

Mar 11, 2014

12060932 (P.T.A.B. Mar. 11, 2014)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________________

BEFORE THE PATENT TRIAL AND APPEAL BOARD
____________________

Ex parte WEI FAN and PHILIP S. YU
____________________

Appeal 2011-009217
Application 12/060,932
Technology Center 2100
____________________

Before BRUCE R. WINSOR, JOHN A. EVANS, and
IRVIN E. BRANCH, Administrative Patent Judges.

BRANCH, Administrative Patent Judge.

DECISION ON APPEAL
Appeal 2011-009217
Application 12/060,932
2
STATEMENT OF CASE
Appellants appeal under 35 U.S.C. § 134(a) from the Examiner’s
Final Rejection of claims 11–15 and 21–35. We have jurisdiction under 35
U.S.C. § 6(b). Claims 1–10 and 16–20 are canceled.
We affirm.
Illustrative Claim
Appellants’ disclosure relates to training an inductive model to output
multiple models from the inductive model, training an error correlation
model to estimate an average output of predictions made by the multiple
models, and determining an error estimation of each of the multiple models
using the error correlation model. Abstract.
Claim 11, reproduced below with the disputed limitations italicized, is
illustrative of the claimed subject matter:
11. A computer-implemented method of diagnosing
prediction model changes, said method comprising:
training an inductive model via a computer;
training an error correlation model via said computer to
estimate an average output of predictions made by said
inductive model; and
determining an error estimation of said inductive model
via said computer using said error correlation model,
wherein said determining of said error estimation
comprises:
supplying a continuous set of variables to said
inductive model;
making predictions with said inductive model
using said continuous set of variables;
producing an average of said predictions;
developing a new feature vector based on said
predictions and said average of said predictions; and
estimating an error of said average of said
predictions.

Appeal 2011-009217
Application 12/060,932
3

Rejections
Claims 11–15 and 21–35 stand rejected under 35 U.S.C. § 102(b) as
being anticipated by “Systematic Data Selection to Mine Concept-Drifting
Data Streams” by Wei Fan (2004) (herein “Fan”). Ans. 3–9.

ANALYSIS
We have reviewed the Examiner’s rejection in light of Appellants’
arguments in the Appeal Brief (“App. Br.” filed Jan 31, 2011) and Reply
Brief (“Reply Br.” filed Apr. 29, 2011). We refer to the Briefs and the
Answer (“Ans.” mailed Apr. 12, 2011) for the respective positions of
Appellants and the Examiner.
Appellants argue claim 11 on the basis that “Fan does not teach or
suggest, ‘supplying a continuous set of variables to said inductive model,’
and ‘making predictions with said inductive model using said continuous set
of variables,’” because Fan describes that “the probability of fraud is a
binary value: either it is true, ‘yes’ or false, ‘no’; not a value within a range
of a continuous dataset.” App. Br. 8, 10 (emphasis omitted); see also App.
Br. 8–12; Reply Br. 3–6. Appellants maintain “[t]he difference a continuous
set of variables has from a binary/categorical set would be that the
continuous set may have an infinite set of variables in it, i.e., ‘... 0, 1, 2, 3, 4,
5, ... ,’ rather than merely a state variable, ‘0/1’ or ‘N/Y’ as described in
Fan.” App. Br. 11; Reply Br. 5.
The Examiner finds Fan’s credit fraud data meets the claimed
“continuous set of variables” supplied to the inductive model and Fan’s
probability of fraud to be the claimed “predictions” made by the inductive
Appeal 2011-009217
Application 12/060,932
4
model using the continuous set of variables. Ans. 5 (referring to Fan §§ 1, 5,
5.1, and 6); see also Ans. 9–16.
Appellants responds that
the transaction data (the “time and money” as identified by the
Examiner) is the transaction data that is processed to determine
the probability of the transaction being a fraud. After all of the
probabilities are computed using stored class distribution
statistics from each decision tree, the probabilities (binary in
nature, having only a “yes” or “no” value), are averaged as the
final membership probability estimate, or input into the
inductive model.

Reply Br. 4.
We are unpersuaded of error. Appellants’ arguments do not persuade
us that Fan’s description of binary probability data negates it describing
“supplying a continuous set of variables to said inductive model” and
“making predictions with said inductive model using said continuous set of
variables.” Appellants do not dispute Fan’s “time and money,” or
transaction data, describes “a continuous set of variables.” Appellants do
not present sufficient persuasive argument or evidence delineating Fan’s
inductive model and therefore have not persuaded us that the transaction
data is supplied to a model input that is not an input to the inductive model.
Moreover, to the extent Appellants argue the claims require
exclusively “supplying a continuous set of variables to said inductive
model,” Appellants’ arguments are not commensurate with the claims and
are therefore unpersuasive. In addition to the transaction data being a
continuous set of variables, Fan describes intermediate values, t(x) (see, e.g.,
Fan § 6 “Credit Card Fraud Data” (calculating t(x))), which also comprise a
continuous set of variables. This continuous set of variable t(x) is then used
to calculate the binary fraud prediction. Id. Accordingly, either the
Appeal 2011-009217
Application 12/060,932
5
transaction data or t(x) meets the claimed “supplying a continuous set of
variables to said inductive model” and “making predictions with said
inductive model using said continuous set of variables.”
On the record before us we are unpersuaded of error in the Examiner’s
rejection of claim 11. Appellants argue claims 21 and 29 on the basis of
arguments presented for claim 11 and, although argued under separate
headings, do not further argue the claims with particularity. Appellants do
not separately argue claims 10–15, 22–28, and 30–35. Accordingly, we
sustain the Examiner’s rejection of claims 11–15 and 21–35.

DECISION1
For the above reasons, the Examiner’s rejection of claims 11–15 and
21–35 is affirmed.
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a)(1)(iv).

AFFIRMED

tj

1 We have decided the appeal before us. The Specification states “[t]he
invention can take the form of . . . an entirely software embodiment . . . .”.
Spec. ¶ 45. In the event of further prosecution, the Examiner should
consider whether claims 11–15 and 21–35 encompass non-statutory subject
matter and should, therefore, be rejected under 35 U.S.C. § 101.