|
Article | Problem | Method | Algorithm | Goal | Extent |
|
[27] | Meaningless time series clustering | Hierarchical and partitioning clustering | -means, hierarchical clustering, EM, SOMS | Proving the claim of meaningless results | No |
|
[45] | Specifying uninteresting sequences and their effects | Density-based clustering | Kernel-density base algorithm | Detecting meaningful pattern | [7, 12] |
|
[78] | Sequential time series clustering is meaningless | Partitioning clustering | -means, distance measuring | Showing sequential time series clustering is not meaningless | [27] |
|
[60] | Very high noise levels | Density-based clustering | Continuous random-walk noise model | Noise elimination and high quality measure | [45] |
|
[14] | Certain constraint in datasets and clusters, meaningless result | Hierarchical and partitioning clustering | Any clustering algorithm | Showing clustering of time series subsequences is meaningless | No |
|
[64] | Reliable determination of the produced sequences of cluster centroids | Partitioning clustering | -means with new distance measure | Results: the claim of the result of -means clustering for time series subsequences is independent of the time series that created it | [14] |
|
[79] | Sinusoidal time series clustering | Partitioning clustering | -means | Explaining sine waves results of subsequence time series clustering | [14] |
|
[5] | Hidden knowledge in time series | Hierarchical clustering, discovery pattern | Adaptive WaveSim transform | Extracting hidden knowledge in time series data | [14] |
|
[28] | Cluster representatives are smoothed and generally do not look at all like any part of the original time series, meaningless results | Hierarchical and partitioning clustering | (Transcription factors) TF-clustering algorithm, TF-minicluster algorithm | Producing useful time series clustering | [27] |
|
[61] | Sequential time series clustering is meaningless | Partitioning clustering | -means clustering by delay vector space | Showing sequential time series clustering can indeed be meaningful | [27] |
|
[13] | Unspecific results from dataset, meaningless | Pattern discovery | RD algorithm | Creating cluster exclusively from subsequences | [14, 60] |
|
[62] | Time consuming to mind the complete set of frequent subsequences for large sequence databases | Pattern discovery | CONTOUR | Efficiently discovering a set of summarization subsequences | No |
|
[72] | Categorizing visitors based on their navigation patterns on a website | Pattern discovery | Repetitive closed gapped subsequence | Constructing feature vector of click stream | [14, 61] |
|
[70] | The detection of repeated subsequences, time series motifs | Pattern discovery | Online motif discovery | Useful extensions of the algorithm to deal with arbitrary data rates and to discover multidimensional motifs. | [75] |
|
[69] | Identifying frequently accurate patterns or motifs | Pattern discovery | Sequitur | Discovery of approximate, variable-length motifs in streaming data. | No |
|