Active Learning for Microarray data
Journal
International Journal of Approximate Reasoning,
Date Issued
2007
Author(s)
DOI
10.1016/j.ijar.2007.03.009
Abstract
In supervised learning it is assumed that it is straightforward to obtain labeled data. However, in reality labeled data
10 can be scarce or expensive to obtain. Active learning (AL) is a way to deal with the above problem by asking for the labels
11 of the most ‘‘informative’’ data points. We propose an AL method based on a metric of classification confidence computed
12 on a feature subset of the original feature space which pertains especially to the large number of dimensions (i.e. examined
13 genes) of microarray experiments. DNA microarray expression experiments permit the systematic study of the correlation
14 of the expression of thousands of genes.
15 Feature selection is critical in the algorithm because it enables faster and more robust retraining of the classifier. The
16 approach that is followed for feature selection is a combination of a variance measure and a genetic algorithm.
17 We have applied the proposed method on DNA microarray data sets with encouraging results. In particular we studied
18 data sets concerning: small round blue cell tumours (four types), Leukemia (two types), lung cancer (two types) and pros-
19 tate cancer (healthy, unhealthy).
10 can be scarce or expensive to obtain. Active learning (AL) is a way to deal with the above problem by asking for the labels
11 of the most ‘‘informative’’ data points. We propose an AL method based on a metric of classification confidence computed
12 on a feature subset of the original feature space which pertains especially to the large number of dimensions (i.e. examined
13 genes) of microarray experiments. DNA microarray expression experiments permit the systematic study of the correlation
14 of the expression of thousands of genes.
15 Feature selection is critical in the algorithm because it enables faster and more robust retraining of the classifier. The
16 approach that is followed for feature selection is a combination of a variance measure and a genetic algorithm.
17 We have applied the proposed method on DNA microarray data sets with encouraging results. In particular we studied
18 data sets concerning: small round blue cell tumours (four types), Leukemia (two types), lung cancer (two types) and pros-
19 tate cancer (healthy, unhealthy).