SAX and Random Projection Algorithms for the Motif Discovery of Orbital Asteroid Resonance Using Big Data Platforms

Area of Science:

• Astronomy
• Data Science
• Computational Science

Background:

• Big data is prevalent in astronomy, notably in analyzing long-term asteroid orbital elements.
• Mean motion resonance (MMR) data is crucial for understanding asteroid-planet interactions over geological timescales.

Purpose of the Study:

• To develop an efficient computational model for quickly obtaining asteroid mean motion resonance (MMR) data.
• To leverage big data platforms for processing large astronomical datasets.

Main Methods:

• Implementation of modified Symbolic Aggregate Approximation (SAX) and motif discovery random projection algorithms.
• Utilizing big data platforms, including Apache Hadoop and Apache Spark, with a five-step process: HDFS storage, RDD import, data preprocessing, UDF-based motif discovery, and result aggregation.
• Development of a computational model integrating SAX and random projection on Hadoop and Spark.

Main Results:

• A significant reduction in computational time compared to standalone methods.
• The proposed model achieved an average accuracy of 83%, validated against SwiftVis software.

Conclusions:

• The big data computational model offers a fast and effective solution for analyzing asteroid orbital elements and identifying MMR.
• This approach demonstrates the power of big data technologies in accelerating astronomical research and data analysis.