The IRMA Community
Newsletters
Research IRM
Click a keyword to search titles using our InfoSci-OnDemand powered search:
|
The Effect of Hidden Units in Neural Networks on Identifying Data Duplication Records
Abstract
Learning algorithms have been widely used to solve different problems in the field of Artificial Intelligence. Presently there are many learning algorithms; each is used depending on specifics of the problem to be solved. Examples of learning algorithms can be found in the field of Artificial Neural Networks (Neural Nets) where these algorithms are used to train the neural nets (as an example, Backpropagation algorithm). Neural nets have been used in data quality problems where a complex database has a lot of duplicate data (dirty data). By using neural nets, it was demonstrated that they can be a very useful tool to identify duplicate and non-duplicate records in the database. In this paper, we show the impact of internal architecture of neural network (hidden units) on the accuracy of results.
|
|