/MOAOnlineClustering

CluStream and DenStream MOA stream clustering algorithms are run with CSV data. Their results, are stored in CSV files. Their configuration is also stored (in a json file), to understand how the parameters influence on the result.

Primary LanguageJava

MOAOnlineClustering

A few MOA stream clustering algorithms are run:

  • CluStream, and
  • DenStream

Both take data from a CSV file, and both resulting clusterings are stored in files.

Note that the algorithm configuration (that meaning parameters values specified before clustering) is also stored, in a json file. That comes in handy when interpreting the results. This "algo config" is stored in the respective algorithm base folder.

❕ Important: configuration

There is a Config class, used to get the paths of the folders in which clustering data will be fetched, and results will be stored. Those paths, have to be specified in a json file. This is done because many applicatios (developed in different languages) use the same folders. Here, a config.json example is shown:

{
    "clusteringResultsPath": "/home/camila/Desktop/TESIS/DATA/clustering_results/",
    "csvDatasetsPath": "/home/camila/Desktop/TESIS/DATA/datasets/",
    "figuresPath": "/home/camila/Desktop/TESIS/DATA/figures/",
    "timeSeriesToyDatasetName": "custom_circumferences_without_k.csv",
    "algoNames": {
	"clustream": "CluStream",
	"denstream": "DenStream"
     }
}

The path to the json file must be specified in the Config class.