Data Mining and Compression

Computaional data mining enables scientisits to discover new scientic knowledge through analyzing, reorgnizing and refining a vast amount of exisiting scientific knowledge as seen in the areas of bioinformatics. Another use of data mining is to argment the accuracy and efficiency of computational methods for scientific and engineering simulations. One of the examples is saving the number of phase-field simualtions for predicting microstructures by mining exisiting microstructure data in repository. To enable such a new approach, we need to develop compact representation of microstructures as well as effective mining scheme. I am currently developing a data compression scheme for microstructures as well as data mining scheme for such a microstructure repository analysis.

Related publications

• Keita Teranishi, Samrat Choudhury, Padma Raghavan, and Long-Qing Chen, Optimization of Parameter Sweeping Simulations for Phase Stability Analysis of Ferroelectric Thin Films. TO be submitted to Jounral of Computer-Aided Materials Design.
• Keita Teranishi, Samrat Choudhury, Tao Wang, Padma Raghavan, Long-Qing Chen and Zi-Kui Liu, Compact data representations of 3-Dimensional Microstructures, To be submitted to Modeling and Simulations in Materials Science and Engineering,
• Keita Teranishi, Tao Wang, Jingxian Zhang, Padma Raghavan, Long-Qing Chen and Zi-Kui Liu, Readily Regenerable Reduced Representation for 2-Dimensional Microstructures, Computational Materials Science, to appear.