StAM Report: Prediction of high risk (explorative) - beta 0.1, 0.9

• Evaluation information
• Annotation information
• Cross validation information
• Model fit information
• Prediction information
• Details on nodes

Information on Evaluation

Total number of samples:	37
Training samples:	37 ( 18 from high risk, 19 from low risk )
Test samples:	37 ( 18 from high risk, 19 from low risk )
Focus on class:	high risk
Data frm chip:	hgu95av2
Start node:	GO:0008150 ( biological_process )

Annotation Information

Chip annotations for:	hgu95av2
Chip data package:	data package version 1.5.1 Mon May 10 11:12:58 2004
GO data package:	GO package version 1.5.1 Mon May 10 11:46:39 2004
Root node chosen:	GO:0008150 ( biological_process )
Number of probesets on the chip:	12625
Probesets annotated below GO:0008150 :	8679
Number of nodes with direct annotations:	1359
Total number of nodes:	2204

Information on Cross Validation

Number of samples:	37
Number of classes:	2
Class weights:	'high risk': 0.1 'low risk': 0.9
Class prior:	'high risk':0.486 'low risk':0.514
Shrinkage candidates:	0, 0.034, 0.069, 0.103, 0.138, 0.172, 0.207, 0.241, 0.276, 0.31, 0.345, 0.379, 0.414, 0.448, 0.483, 0.517, 0.552, 0.586, 0.621, 0.655, 0.69, 0.724, 0.759, 0.793, 0.828, 0.862, 0.897, 0.931, 0.966, 1

Cross validated performance

Download postscript

Download postscript

	Root.error.rate	Root.performance	Mean.redundancy	Nodes	Genes
0	0.16216	1.03954	0.70034	1869	8580
0.034	0.16216	0.98103	0.7096	1221	8325
0.069	0.13514	0.9268	0.72645	798	8082
0.103	0.13514	0.87855	0.74504	600	7489
0.138	0.13514	0.84157	0.76923	460	6939
0.172	0.13514	0.81505	0.81423	335	6730
0.207	0.13514	0.79004	0.82966	304	6609
0.241	0.13514	0.76166	0.86688	247	5958
0.276	0.10811	0.73314	0.89929	206	5865
0.31	0.10811	0.71119	1.01677	145	5482
0.345	0.10811	0.69594	1.10051	122	5176
0.379	0.10811	0.68495	1.17571	100	4839
0.414	0.10811	0.68387	1.19868	89	4592
0.448	0.10811	0.67754	1.25239	76	4495
0.483	0.16216	0.66688	1.39388	60	4142
0.517	0.16216	0.65293	1.58789	49	3917
0.552	0.16216	0.64222	1.90103	39	3464
0.586	0.16216	0.63613	1.92392	39	3464
0.621	0.16216	0.63155	1.87321	36	3192
0.655	0.10811	0.61515	1.79705	34	3120
0.69	0.13514	0.56581	2.31609	21	2626
0.724	0.13514	0.51405	2.21575	21	2626
0.759	0.05405	0.48077	3.02657	15	1687
0.793	0.05405	0.47508	2.4398	15	1687
0.828	0.05405	0.48382	1.98035	15	1687
0.862	0.05405	0.488	2.77233	9	561
0.897	0.05405	0.49602	2.79196	9	561
0.931	0.05405	0.49075	2.7913	9	561
0.966	0.05405	0.46984	2.78983	9	561

Information on Model Fit

Performance vs. redundancy

Download postscript

	Best Delta	Root Error Rate	Performance	Redundancy	#Nodes	#Genes
0	0	0.1622	1.0395	0.7003	1869	8580
0.1	0.1035	0.1351	0.8786	0.745	600	7489
0.2	0.138	0.1351	0.8416	0.7692	460	6939
0.3	0.2759	0.1081	0.7331	0.8993	206	5865
0.4	0.2759	0.1081	0.7331	0.8993	206	5865
0.5	0.2759	0.1081	0.7331	0.8993	206	5865
0.6	0.8278	0.0541	0.4838	1.9804	15	1687
0.7	0.8278	0.0541	0.4838	1.9804	15	1687
0.8	0.8278	0.0541	0.4838	1.9804	15	1687
0.9	0.8278	0.0541	0.4838	1.9804	15	1687
1	0.9658	0.0541	0.4698	2.7898	9	561

Fit for D = 0.8278

Number of classes:	2
Class weights:	'high risk': 0.1 'low risk': 0.9
Class prior:	'high risk':0.486 'low risk':0.514
Chosen shrinkage level:	0.828 (by accepting the system default (after consulting weighting alternatives) )
Number of nodes left	15
Number of accessible genes	1687

Shrunken prediction graph

Hover over a node to see its description, click on it to view its detailed results.

Download postscript

Nodewise evaluation

Download postscript

Information on Predictions

Number of samples:	37 (18 from high risk, 19 from low risk)
Classes:	high risk, low risk

Illustration of Performance on testdata

Nodewise evaluation

Download postscript

Prediction image

Hover over a node description to see its ID, click on it to see detailed results.

Download postscript

Details on Nodes

GO:0008150

biological_process

Biological Process

Parent node

GO:0003673	Gene_Ontology (not in fit)

8 child nodes

ID	Term	Weight
GO:0007610	behavior	0.00000
GO:0000004	biological_process unknown	0.00000
GO:0009987	cellular process	0.00000
GO:0007275	development	0.00000
GO:0008371	obsolete biological process	0.00000
GO:0007582	physiological process	1
GO:0050789	regulation of biological process	0.00000
GO:0016032	viral life cycle	0.00000

Performance on training data

• Error rate: 0.08108
• Performance: 0.48382
• Redundancy: 2.20323

	Sensitivity	Specificity
high risk	0.94444	0.89474
low risk	0.89474	0.94444

Performance on test data

• Error rate: 0.08108
• Performance: 0.34471
• Redundancy: 2.85707

	Sensitivity	Specificity
high risk	0.94444	0.89474
low risk	0.89474	0.94444

Predictions

	Class	high risk	low risk	Predictionn
00290952H95A2T01407 (testset)	low risk	0.06403	0.93597	low risk
00290964H95A2T01439 (testset)	low risk	0.127	0.873	low risk
00290966H95A2T01461 (testset)	low risk	0.00832	0.99168	low risk
00291004H95A2T01194 (testset)	high risk	0.53919	0.46081	high risk
00291005H95A2T01047 (testset)	low risk	0.00017	0.99983	low risk
00291046HU95A2-T01-502 (testset)	high risk	0.93334	0.06666	high risk
00291048HU95A2-T01-514 (testset)	low risk	0.01058	0.98942	low risk
00291053HU95A2-T01-618 (testset)	low risk	0.02301	0.97699	low risk
00291087H95A2-T01-255 (testset)	high risk	0.99704	0.00296	high risk
00291092H95A2-T01-838 (testset)	high risk	0.99375	0.00625	high risk
00291126H95A2-T010310 (testset)	high risk	0.87897	0.12103	high risk
00291127H95A2-T010585 (testset)	high risk	0.58826	0.41174	high risk
00291165HU95A2T01-472 (testset)	high risk	0.9981	0.0019	high risk
00291187H95A2-T00-0064 (testset)	high risk	0.9919	0.0081	high risk
00291188H95A2-T00-0070 (testset)	high risk	0.99606	0.00394	high risk
00291189H95A2-T00-0109 (testset)	high risk	0.99949	0.00051	high risk
00291191H95A2-KF67 (testset)	high risk	0.84325	0.15675	high risk
0029_1208_H95A2_KF0076 (testset)	low risk	0.12226	0.87774	low risk
0029_1209_H95Av2_KF0077 (testset)	low risk	0.00174	0.99826	low risk
0029_1210_H95A2_KF0079 (testset)	low risk	0.00515	0.99485	low risk
0029_1213_H95A2_KF0110 (testset)	low risk	0.58274	0.41726	high risk
0029_1221_H95A2_KF0144 (testset)	low risk	0.42142	0.57858	low risk
0029_1225_H95A2_KF0157 (testset)	low risk	0.20175	0.79825	low risk
0029_1264_H95A2_KF145 (testset)	high risk	0.49428	0.50572	low risk
0029_1265_H95A2_KF151 (testset)	high risk	0.86309	0.13691	high risk
0029_1266_H95A2_KF155 (testset)	high risk	0.9206	0.0794	high risk
0029_1287_H95A2_KF-184 (testset)	high risk	0.9381	0.0619	high risk
0029_1289_H95A2_KF-201 (testset)	high risk	0.82717	0.17283	high risk
0029_1291_H95A2_KF-097 (testset)	high risk	0.89811	0.10189	high risk
0029_1340_H95A2_KF-069 (testset)	high risk	0.97824	0.02176	high risk
0029_1349_H95A2_KF-113 (testset)	low risk	0.02371	0.97629	low risk
0029_1351_H95A2_KF-117 (testset)	low risk	0.20922	0.79078	low risk
0029_1352_H95A2_KF-120 (testset)	low risk	0.00313	0.99687	low risk
0029_1375_H95A2_KF-094 (testset)	low risk	0.13331	0.86669	low risk
0029_1376_H95A2_KF-095 (testset)	low risk	0.12915	0.87085	low risk
0029_1377_H95A2_KF-104 (testset)	low risk	0.5106	0.4894	high risk
0029_1396_H95A2_KF-105 (testset)	low risk	0.0171	0.9829	low risk