| |
|
|
| |
Proprietary
Technology
Highlights
|
|
| |
|
|
| |
-
“Composite
testing”
for
increased
data
consideration
for
models.
-
Opportunistic
data
pooling
technique
for
improved
model
building.
|
|
|
|
|
-
Entropy
based
measures
(from
DSP)
combined
with K-S
metrics
from
statistical
separability
analysis.
-
Clustering
techniques
for
segmentation
that are
specifically
geared
towards
predictive
modeling
with
neural
networks.
|
|
| |
-
Variable
representation
techniques
for
different
types of
variables,
uniquely
designed for
real world
neural
network
implementation.
-
Unique
method for
“reject
inference”
and custom
sampling of
data for
model
building.
|
|
| |
|
| |
|
