Mining information and knowledge from very large databases is recognized as a key research area in the machine learning and expert systems field. We show connectionist models can be used to learn software effort when the available data set is small. We use a connectionist model to learn software effort from a small training data set of several software projects and validate the model on a holdout sample. Several design issues associated with developing connectionist models on a small training data set are described. Our research indicates that connectionist models, whenever carefully designed, could help knowledge discovery and software effort estimation.
|Original language||English (US)|
|Number of pages||8|
|Journal||Journal of Computer Information Systems|
|State||Published - Sep 2002|
All Science Journal Classification (ASJC) codes
- Information Systems