Intelligent Engineering Systems through Artificial Neural Networks, Volume 20
78 Committee Network Model for HDD Functional Tests
Download citation file:
In this paper, a committee network model was used as a substitute for HDD functional test procedures, where a particular drive required approximately more than ten hours to perform the test. Here, the committee network model classified HDDs into either “pass” or “fail” within a couple milliseconds. Thus, if this model can be used to replace the traditional test procedures, significant amount of cycle time can be reduced. This paper demonstrated a committee network construction, where the training data for the patterns of defective HDDs was sparse compared to the good patterns. Our paper also developed a method to deal with sparseness of the failed patterns of HDDs. Experimental results suggested the promising outcome that the committee network can be used to tame the problem of sparse data and thus the model can potentially be used as a substitute for the HDD functional tests.