Loading...
Please wait, while we are loading the content...
Similar Documents
A Novel K means Clustering Algorithm for Large Datasets Based on Divide and Conquer Technique
| Content Provider | Semantic Scholar |
|---|---|
| Author | Ahirwar, Rajesh B. |
| Copyright Year | 2014 |
| Abstract | In this paper we propose an efficient algorithm that is based on divide and conquers technique for clustering the large datasets. In our research work we have applied divide and conquer technique on partitions of the large datasets and we have used squared Euclidean distance for measuring the similarity between data points. The partitioning of datasets is done according to the number of clusters desired. Finally clusters are obtained from each partition of the dataset and we merge those clusters to get more precise clusters. Our proposed technique uses two phases with seven steps for clustering the large datasets. The advantage of using divide and conquer technique is that the large datasets which require a large amount of physical memory to load into the system can also be clustered using our proposed algorithm as it requires a small amount of physical memory because the clustering is done on parts of the dataset. Finally we have used three performance measures namely Fmeasure, purity and entropy to compare our results from the existing algorithms. The results have shown that our approach is much better than existing algorithms. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://www.ijcsit.com/docs/Volume%205/vol5issue01/ijcsit2014050163.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
