Computer Science – Learning
Scientific paper
Dec 2009
adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=2009agufmsm53c..02k&link_type=abstract
American Geophysical Union, Fall Meeting 2009, abstract #SM53C-02
Computer Science
Learning
[1914] Informatics / Data Mining, [1942] Informatics / Machine Learning, [2722] Magnetospheric Physics / Forecasting, [2774] Magnetospheric Physics / Radiation Belts
Scientific paper
Heliophysics is a data centric field which relies heavily on the use of spacecraft data for further advances. The prevalent approach to analysis of spacecraft data is based on visual inspection of data. As a result, the vast majority of the collected data from various missions has gone unexplored. The computer aided algorithmic approach to data analysis as facilitated through data mining techniques are essential for analysis of large data sets and enable discovery of hidden information and patterns in the data. Many data analysis problems in space weather stand to benefit from the application of data mining techniques. Examples include identifying spacecraft charging signatures in plasma detectors, identifying plasma frequency lines in wave spectrograms (and hence density), detecting and classifying substorm infection features, among others (R. Friedel, private communication). Thus while the need for advanced algorithmic approach to data exploration and knowledge discovery is generally recognized by experimentalists, the adoption of such techniques (“data mining”) has been slow. This has been partly due to the steep learning curve of some of the techniques and/or the requirement to have a working knowledge of statistics. Another factor is the existence of a plethora of data mining approaches, and it is often a daunting task for a scientist to determine the appropriate technique. Our goal has been to make such tools accessible to non-experts and remove it from gee-whiz domain to a practical tool that will become part of the standard arsenal of data analysis. To this end, we have developed an automated data mining technique called MineTool. Its first deployment to analysis of Cluster has been very successful (Karimabadi et al., JGR, 114, A06216 , 2009) and this tool is gaining adoption among experimentalists. In this talk, we will provide an overview of this tool, illustrate its use through examples, and discuss future directions of research.
Karimabadi Homa
Sipes T.
No associations
LandOfFree
Automated Data Preparation and Physics Mining Tools for Space Weather Studies (Invited) does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with Automated Data Preparation and Physics Mining Tools for Space Weather Studies (Invited), we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Automated Data Preparation and Physics Mining Tools for Space Weather Studies (Invited) will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-1881786