Abstract
Too huge amount of data would be a strict challenge to any kind of data mining techniques and algorithms. This study introduces the concept of information granulation to reduce the data size. We use Fuzzy ART to construct information granules and extract knowledge from them. The indexes, homogeneity and undistinguishable ratio are developed to determine the granularity of information. The proposed approach does not only decrease the amount of data, but also maintain the characteristics of data.