| Due to advances in communication and technology, a significant area of data mining has emerged, mining data streams. Clustering, a traditional method used in knowledge discovery, uncovers patterns in data that were not previously known. In the recent years, the amount of data received and collected by organizations has more than quadrupled, hence the need for faster, precise and memory efficient algorithms.; In this study, an algorithm that uses DCT transformed data is presented. This is a grid and density-based clustering algorithm that uses an estimated distribution of the data stream to find arbitrary shaped clusters.; The experimental results indicate that the clustering algorithm provides useful results and confirms that DCT is a very reliable tool for compressing data streams. The patterns in the data stream distribution are well-preserved after reconstruction even when very few of the coefficients are taken into consideration. |
