HierArchical-Grid CluStering Based on DaTA Field in Time-Series and the Influence of the First-Order Partial Derivative Potential Value for the ARIMA-Model
Extend the function of static-time dataframe clustering algorithm (HASTA: HierArchical-grid cluStering based on daTA field) to be able to cluster the time-series dataframe. The algorithm purposed to use a set of “first-partial derivative potential value” given from HASTA in the multiple dataframes as the input to the autoregressive integrated moving average (ARIMA) under preliminary parameters. The ARIMA model could perform the pre-labeling task for the cluster(s) in the connected dataframe on the same time-series data. Calculating the structural similarity as a distance measure between timeframe, ARIMA would mark the high potential grid(s) as the cluster tracker. As the result, the ARIMA model could interpreting and reasoning cluster movement phenomena in the systematic approach. This integration is the attempt to show the influent power of data field in the term of knowledge representation.
KeywordsKnowledge discovery Pattern recognition Data field Time-series Time-series clustering ARIMA
This work was supported by the National Key Research and Development Program of China (No.2016YFB0502600), the National Natural Science Fund of China (61472039), Beijing Institute of Technology International Cooperation Project (GZ2016085103), and Open Fund of Key Laboratory for National Geographic Census and Monitoring, National Administration of Surveying, Mapping and Geoformation (2017NGCMZD03).
- 2.Banaezadeh, F.: ARIMA-modeling based prediction mechanism in object tracking sensor networks. In: 2015 7th Conference on Information and Knowledge Technology (IKT), pp. 1–5 (2015)Google Scholar
- 3.Fränti, P.: Clustering basic benchmark (2015)Google Scholar
- 6.Li, D., Wang, S., Yuan, H., Li, D.: Software and applications of spatial data mining. Wiley Interdiscip. Rev.: Data Min. Knowl. Discov. (Wiley Online Libr.) 6, 84–114 (2016)Google Scholar
- 8.Sardá-Espinosa, A.: Comparing time-series clustering algorithms in R using the dtwclust Package. Accessed May 2017Google Scholar
- 11.Press, W.H., Teukolsky, S.A., Vetterling, W.T., Flannery, B.P.: Gaussian mixture models and k-means clustering. In: Numerical Recipes: The Art of Scientific Computing. Cambridge University Press, New York (2007)Google Scholar
- 13.Yan, Z.: Traj-ARIMA: a spatial-time series model for network-constrained trajectory. In: Proceedings of the Third International Workshop on Computational Transportation Science, pp. 11–16 (2010)Google Scholar