_PROBLEM CoEPrA-2006_Classification_001


_GROUP_NAME Scott Oloff

_GROUP_MEMBERS
Scott Oloff

_ADDRESS
Boehringer Ingelheim
900 Ridgebury Rd.
Ridgefield, CT 06877

_MODELING_PROCEDURE
I used a newly developed machine learning approach called kScore, which
is based on kNN and SVM technologies.  The approach is kernel free and
uses all training compounds in a distance-weighted manner to predict a
test compound rather than just its k Nearest Neighbors.  The algorithm
also incorporates Generalization constraints, a soft margin classifier
(for Classification), and an Epsilon Loss function (for Regression)
similar to those proposed by V. Vapnik.  There is a built in
Applicability Domain for each kScore model and a more detailed
explanation of the method is being submitted for publication.

The descriptors of the original training set were normalized from 0 to
1.  The training set was then randomly divided into 5 training and test
sets whereby the test set contained 20% of the compounds.  For each
training/test split several kScore models were generated (using the
COEPRA provided descriptors) to find the best values of C and Epsilon
for external prediction.  The most predictive models were collected and
used in a consensus fashion to predict the external test set.  The
average LOO Training accuracies were 0.9-0.95 and the average Test
accuracies for each class varied between 0.8 and 1.0.  The most
predictive values of Epsilon and C varied from 0.6 to 0.8 and 500 to
1000, respectively.  The descriptor weights of the most predictive 
model
are attached.

The predictions with the lowest confidence are:
Obj_00002
Obj_00019
Obj_00023
Obj_00028
Obj_00041
Obj_00052
Obj_00057
Obj_00059
Obj_00060
Obj_00068
Obj_00078

_PREDICTION
Obj_00001 -1
Obj_00002 -1
Obj_00003 -1
Obj_00004 -1
Obj_00005 +1
Obj_00006 +1
Obj_00007 -1
Obj_00008 -1
Obj_00009 -1
Obj_00010 +1
Obj_00011 +1
Obj_00012 +1
Obj_00013 +1
Obj_00014 +1
Obj_00015 -1
Obj_00016 -1
Obj_00017 -1
Obj_00018 +1
Obj_00019 -1
Obj_00020 -1
Obj_00021 +1
Obj_00022 +1
Obj_00023 +1
Obj_00024 +1
Obj_00025 +1
Obj_00026 +1
Obj_00027 +1
Obj_00028 -1
Obj_00029 +1
Obj_00030 +1
Obj_00031 +1
Obj_00032 +1
Obj_00033 -1
Obj_00034 -1
Obj_00035 +1
Obj_00036 -1
Obj_00037 -1
Obj_00038 -1
Obj_00039 -1
Obj_00040 +1
Obj_00041 -1
Obj_00042 -1
Obj_00043 -1
Obj_00044 +1
Obj_00045 +1
Obj_00046 -1
Obj_00047 +1
Obj_00048 +1
Obj_00049 +1
Obj_00050 +1
Obj_00051 +1
Obj_00052 +1
Obj_00053 -1
Obj_00054 -1
Obj_00055 +1
Obj_00056 +1
Obj_00057 -1
Obj_00058 -1
Obj_00059 +1
Obj_00060 +1
Obj_00061 -1
Obj_00062 +1
Obj_00063 +1
Obj_00064 -1
Obj_00065 +1
Obj_00066 +1
Obj_00067 +1
Obj_00068 -1
Obj_00069 +1
Obj_00070 -1
Obj_00071 -1
Obj_00072 +1
Obj_00073 -1
Obj_00074 -1
Obj_00075 -1
Obj_00076 -1
Obj_00077 -1
Obj_00078 +1
Obj_00079 -1
Obj_00080 +1
Obj_00081 +1
Obj_00082 -1
Obj_00083 -1
Obj_00084 +1
Obj_00085 -1
Obj_00086 -1
Obj_00087 +1
Obj_00088 -1