Asteroid Track Candidate Construction by Distributed Sequential Pattern Mining on Large Astronomical Dataset

With the recent advance of astronomical observation technology, the volume of collected data hasaccumulated at size of Petabytes level. It is inevitable to apply Big-data approach to analyze themassive datasets.Quite a few projects have constructed advanced telescopes and use them to survey various spaceobjects, including asteroid movements. The observed datasets are then analyzed by astronomicalresearchers by various ways. The goal of this research is to develop methods for analyze asteroids trackcandidate with better accuracy, by applying distributed algorithm on cloud environments. Large setsof analyzed results and intermediate data are systematically maintained in our design, and thesedatasets can be used for following projects’ observation data. We will also discuss the performanceissue with parameters concerning cloud environments.After constant discussions with astronomical researchers, in this proposal we emphasize ongenerating track candidates complied with the observation sequence. This requirement is enforced tofurther filtering out candidates in our previous design. The more accurate and concise results will bederived by distributed sequential pattern analysis algorithm. This methodology may benefit many otherapplications with sequential analysis concerns.