1 / 16

Different types of Spatio -temporal Data Mining

Different types of Spatio -temporal Data Mining. By Mukund Malladi Rajasekhar Ganduri Siddhartha Katragadda. Contents. Introduction Related Work Experiment Results Comparison of Papers Conclusion References. Introduction.

nara
Télécharger la présentation

Different types of Spatio -temporal Data Mining

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Different types of Spatio-temporal Data Mining By Mukund Malladi Rajasekhar Ganduri Siddhartha Katragadda

  2. Contents • Introduction • Related Work • Experiment • Results • Comparison of Papers • Conclusion • References

  3. Introduction • All three paper deals with proposing new data mining techniques for analyzing information in large Spatio-Temporal Databases. • In paper-1 ,the cluster analysis collected of air pollutants are used in identifying spatial variation with time which can useful in adapting new procedures for controlling it. • The second paper deals with mining spatial and temporal data relation in scientific data, in which both geometric properties of features and techniques to model spatial relationships among features are main area. • In third paper a new system is proposed which contains new technology to efficiently support data-mining process ,address spatial and temporal dimensions of dataset and visualize and interpret results

  4. Related Work • The paper-1 uses multi-scale wavelet transformation instead of Short –Time-Fourier Transform. • It proposed two techniques called DWT and CWT for analyzing the spatio-temporal data. • Self-organizing Map (SOM)is also proposed which helps in providing informative pictures of the data space and also comparing datasets with each other.

  5. Cont. Mining spatial and spatio-temporal relation- ships in scientific data is also very challenging. • Geometric properties of features – Robust Technique • Techniques to model spatial relationships among features – Fast algorithms for extracting frequent data • one needs to develop effective approaches to incorporate temporal information into the overall analysis– SOAP approach • effective reasoning methods are needed to make inferences on important events, such as defect amalgamation in materials, based on the extracted spatial and spatio-temporal relationships – Empirically evaluated on real case study applications

  6. Contd. • Due to the huge volume and diverse nature of this kind of data, traditional techniques such as statistical methods are inadequate to explain complex spatial and temporal relationships among data. • The author uses clustering and association rules for this purpose. • Approach of this paper • Localiser – deals with spatial and temporal dimensions • Miner - processes the data based on the spatiotemporal relationships provided by the localiser.

  7. Experiment • For paper-1 the data is taken from Taiwan Air Quality Monitoring Network(8 stations) .At first data cleaning is done followed by data transformation(DWT and CWT). • Mining techniques are used like AugSOM a JAVA enabled tool. • The cohesion measure of a cluster which indicates the similarity between a pair of stations in the cluster are determined and also variance is determined. • For different number of clusters the ,DWT and CWT is performed and the clustering performance is recorded. • The data is recorded daily ,weekly ,monthly and yearly then P and V are calculated.

  8. Experiment Cont. • Spatial Feature Representation : Three main different representation schemes • Parallelepiped – for regular shapes • Ellipse - for vortices • Landmarks – for irregular shapes • SOAP(Spatial Object Association Pattern) characterizes “closeTo” or “isAbove” relationships among Star (closeTo), Clique(closeTo), Sequence(isAbove), minlink(user specified). • Framework:

  9. SOAP Mining: • Data Organization: reduces the search space by eliminating of infrequent SOAPs. • Equivalence Classes: algorithm can efficiently discover different types of SOAPs from large amounts of data. • Anti-monotone Property: a set of objects are frequent only if, all of their subsets are frequent , to reduce the search space. • Optimizations: strategies to quickly identify valid neighbors of an object.

  10. Contd. Visual Techniques for Spatial Analysis : • Visual Data Mining refers to methods, approaches and tools for the exploration of large datasets by allowing users to directly interact with visual representations of data and dynamically modify parameters to see how they affect the visualized data. • The visualization tool displays in 3D the Association Rules identified by the mining engine. • The tool exploits the “Arranging view” visualization technique, where two different views are presented in separated windows, and the user can arbitrarily arrange them to facilitate the comparison of data.

  11. Conclusion • In paper -1 the ,cluster analysis is performed data is analysed from multi-scales CWT and DWT. It shows that regions determined from the wavelet transform approach reduces local small regions using the small scale input data and improve the over-smoothed regions using one large scale input data. • SOM, an excellent visual tool for investigating the inner structure of the transformed data, provides the capability for hard clustering.

  12. Contd. • Framework for mining spatial and spatio-temporal patterns in scientific data sets. • models features as geometric objects rather than points. • supports multiple distance measurements. • accommodated temporal information in the overall analysis to characterize the evolutionary behavior of an interaction.

  13. Contd. • In this the author introduced a visualization tool for viewing and interacting with the results of data mining. • The association rules are used for investigating different areas for antecedent and consequent.

  14. Comparisons:

  15. References: • [1] Sheng-Tun Li ;Shih-Wei Chou ,Multi-Resolution Spatio-temporal Data Mining for the Study of Air Pollutant Regionalization , In proceedings of 33th Hawaii International Conference on System Science-200, Page(s): 1-7, 2000. • [2] Hui Yang ;SrinivasanParthasarathy ,Mining Spatial and Spatio-temporal Patterns in Scienctific Data, Knowledge and Data Engineering, IEEE Transactions on , Page(s): 433 – 448,2008. • [3] M-TaharKechadi; MichelaBertolotto, A Visual Approach for Spatio-Temporal Data Mining, Information Reuse and Integration, 2006 IEEE International Conference on, Page(s): 504 - 509,2006.

  16. Thank you

More Related