Name
Abstract | ||
|---|---|---|
•We evaluated eight different concept drift detectors.•A 2k factorial design was used to indicate the best parameters for each method.•Tests compared accuracy, evaluation time, false alarm and miss detection rates.•A Mahalanobis distance is proposed as a metric to compare drift methods.•DDM was the method that presented the best average results in all tested datasets. |
Year | DOI | Venue |
|---|---|---|
2014 | 10.1016/j.eswa.2014.07.019 | Expert Systems with Applications |
Keywords | Field | DocType |
Data streams,Time-changing data,Concept drift detectors,Comparison | Data mining,Data stream mining,False alarm,Computer science,Data stream,Mahalanobis distance,Concept drift,Survey research,Artificial intelligence,Detector,Machine learning,Statistical hypothesis testing | Journal |
Volume | Issue | ISSN |
41 | 18 | 0957-4174 |
Citations | PageRank | References |
11 | 0.53 | 25 |
Authors | ||
4 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Paulo Mauricio Gonçalves Jr. | 1 | 11 | 0.53 |
| Silas Garrido Teixeira de Carvalho Santos | 2 | 54 | 5.01 |
| Roberto Souto Maior de Barros | 3 | 19 | 1.07 |
| Davi Carnauba De Lima Vieira | 4 | 11 | 0.53 |