2009
Woodring, Jonathan; Shen, Han-Wei
Semi-Automatic Time-Series Transfer Functions via Temporal Clustering and Sequencing Proceedings Article
In: Computer Graphics Forum, pp. 791–798, Wiley Online Library 2009.
Abstract | Links | BibTeX | Tags: temporal clustering and Sequencing, time-series transfer functions
@inproceedings{woodring2009semi,
title = {Semi-Automatic Time-Series Transfer Functions via Temporal Clustering and Sequencing},
author = {Jonathan Woodring and Han-Wei Shen },
url = {http://datascience.dsscale.org/wp-content/uploads/2016/06/Semi-AutomaticTime-SeriesTransferFunctionsViaTemporalClusteringAndSequencing.pdf},
year = {2009},
date = {2009-01-01},
booktitle = {Computer Graphics Forum},
volume = {28},
number = {3},
pages = {791--798},
organization = {Wiley Online Library},
abstract = {When creating transfer functions for time-varying data, it is not clear what range of values to use for classification, as data value ranges and distributions change over time. In order to generate time-varying transfer functions, we search the data for classes that have similar behavior over time, assuming that data points that behave similarly belong to the same feature. We utilize a method we call temporal clustering and sequencing to find dynamic features in value space and create a corresponding transfer function. First, clustering finds groups of data points that have the same value space activity over time. Then, sequencing derives a progression of clusters over time, creating chains that follow value distribution changes. Finally, the cluster sequences are used to create transfer functions, as sequences describe the value range distributions over time in a data set.},
keywords = {temporal clustering and Sequencing, time-series transfer functions},
pubstate = {published},
tppubtype = {inproceedings}
}
When creating transfer functions for time-varying data, it is not clear what range of values to use for classification, as data value ranges and distributions change over time. In order to generate time-varying transfer functions, we search the data for classes that have similar behavior over time, assuming that data points that behave similarly belong to the same feature. We utilize a method we call temporal clustering and sequencing to find dynamic features in value space and create a corresponding transfer function. First, clustering finds groups of data points that have the same value space activity over time. Then, sequencing derives a progression of clusters over time, creating chains that follow value distribution changes. Finally, the cluster sequences are used to create transfer functions, as sequences describe the value range distributions over time in a data set.
: . .
1.
Woodring, Jonathan; Shen, Han-Wei
Semi-Automatic Time-Series Transfer Functions via Temporal Clustering and Sequencing Proceedings Article
In: Computer Graphics Forum, pp. 791–798, Wiley Online Library 2009.
@inproceedings{woodring2009semi,
title = {Semi-Automatic Time-Series Transfer Functions via Temporal Clustering and Sequencing},
author = {Jonathan Woodring and Han-Wei Shen },
url = {http://datascience.dsscale.org/wp-content/uploads/2016/06/Semi-AutomaticTime-SeriesTransferFunctionsViaTemporalClusteringAndSequencing.pdf},
year = {2009},
date = {2009-01-01},
booktitle = {Computer Graphics Forum},
volume = {28},
number = {3},
pages = {791--798},
organization = {Wiley Online Library},
abstract = {When creating transfer functions for time-varying data, it is not clear what range of values to use for classification, as data value ranges and distributions change over time. In order to generate time-varying transfer functions, we search the data for classes that have similar behavior over time, assuming that data points that behave similarly belong to the same feature. We utilize a method we call temporal clustering and sequencing to find dynamic features in value space and create a corresponding transfer function. First, clustering finds groups of data points that have the same value space activity over time. Then, sequencing derives a progression of clusters over time, creating chains that follow value distribution changes. Finally, the cluster sequences are used to create transfer functions, as sequences describe the value range distributions over time in a data set.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
When creating transfer functions for time-varying data, it is not clear what range of values to use for classification, as data value ranges and distributions change over time. In order to generate time-varying transfer functions, we search the data for classes that have similar behavior over time, assuming that data points that behave similarly belong to the same feature. We utilize a method we call temporal clustering and sequencing to find dynamic features in value space and create a corresponding transfer function. First, clustering finds groups of data points that have the same value space activity over time. Then, sequencing derives a progression of clusters over time, creating chains that follow value distribution changes. Finally, the cluster sequences are used to create transfer functions, as sequences describe the value range distributions over time in a data set.