在科學的領域中，天文學具有很重要的地位。由於近年來觀測技術及硬體設備不斷提升，讓天文領域的研究者，能進行更多樣化的分析，而天文望遠鏡所觀測的天文數據資料，歷日曠久不斷累積增加，數據量已逐漸增長到PB（Petabyte）等級的巨量資料（Big Data）。面對單機系統無法負荷的巨大數據量，需使用分散式運算，才能夠有效加速處理分析的運算時間。
本論文提出了基於Hadoop的分散式架構下，用以協助天文學者分類變星（variable stars）星體的字尾樹系統，系統使用MapReduce與Spark兩種分散式運算框架設計，系統在建構字尾樹的階段，是將大量星體隨時間改變亮度的序列資料，以字尾樹的形式轉成樹狀結構，儲存至分散式的檔案系統中，並支援對於後續資料的新增。利用字尾樹的特性，能讓使用者進行高效率的查詢，此外，系統的檢索階段引入了階層化（Hierarchical）的概念，能夠調整樹中資料的細膩程度，除了能找出因觀測或計算誤差產生的類似序列，亦能夠因應不同的分類方式，提供更宏觀的查詢，讓天文研究者在分類星體時，能依照不同的需求選擇相應的細膩度，來快速地找到，擁有相同或是相似特徵的星體編號。

In the field of science, astronomy has a very important status. As the observation technology and hardware equipment in recent years continue to improve, so that researchers in the field of astronomy can do more diversified analysis, and the amount of data observed by astronomical telescope continue to increase, and has gradually increased to Petabyte level.
In this paper, a suffix tree system based on distributed sturcture of Hadoop is proposed to assist astronomers to classify variable stars. The system is designed with MapReduce and Spark frameworks. In the stage of constructing suffix tree, the system converts a large amount of data, which is the sequence of star brightness changing over time, into a suffix tree structure, then stores the tree in the distributed file system; the system also supports appending following observation data. Using the characteristics of the suffix tree allows users to query efficiently. Moreover, the query stage of the system introduces the hierarchical concept, which can adjust the preciseness of the data in the tree, allows the system to not only find out the similar sequence generated by observation or calculation errors but also provide more diversified query in response to different classification methods. According to different needs, astronomical researchers can select the preciseness of data to classify stars, and quickly find the ID of same or similar characteristics of the star.

[1] Pan-STARRS, http://pan-starrs.ifa.hawaii.edu/public/
[2] 陳文屏, 「天文觀測的新挑戰─談泛星計畫」, 科儀新知, 第30卷第3期, 2008.
[3] Wikipedia, “variable star”, https://en.wikipedia.org/wiki/Variable_star
[4] Apache Hadoop, http://hadoop.apache.org/
[5] Konstantin Shvachko, Hairong Kuang, Sanjay Radia, Robert Chansler, “The Hadoop Distributed File System,” MSST, 2010.
[6] HDFS, https://hadoop.apache.org/docs/r1.2.1/hdfs_design.html
[7] Jeffrey Dean and Sanjay Ghemawat, “MapReduce: Simplified Data Processing on Large Clusters” OSDI'04: Sixth Symposium on Operating System Design and Implementation,San Francisco, CA, December, 2004.
[8] Hadoop 101: Programming MapReduce with Native Libraries, Hive, Pig, and Cascading, http://blog.pivotal.io/pivotal/products/hadoop-101-programming-mapreduce-with-native-libraries-hive-pig-and-cascading
[9] Apache Spark, https://spark.apache.org/
[10] Spark Cluster, https://spark.apache.org/docs/latest/cluster-overview.html
[11] Matei Zaharia, Mosharaf Chowdhury, Tathagata Das, Ankur Dave, Justin Ma, Murphy McCauley, Michael J. Franklin, Scott Shenker, Ion Stoica, “Resilient Distributed Datasets: A Fault-Tolerant Abstraction for In-Memory Cluster Computing,” In NSDI, 2012.
[12] P. Weiner, ”Linear Pattern Matching Algorithm,” 14th Annual IEEE Symposium on Switching and Automata Theory, 1973.
[13] Min-Feng Wang, Chi-Sheng Huang*, Meng-Feng Tsai, Bo-Ru Song, Shin-Fu Su and Cheng-Hsien Tang, “Generalized Analysis of Message Propagation on Social Network,” International Journal of Future Generation Communication and Networking Vol. 5, No. 2, June, 2012.
[14] 沈敬軒, “Mining Similar Astronomical Sequence Pattern with Hierarchical Weighted Suffix Tree,” 國立中央大學, 碩士論文, 2011.
[15] 張哲嘉, “Distributed Suffix Tree Based Sequential Pattern Management System for Astronomical Analysis,” 國立中央大學, 碩士論文, 2013.
[16] 蔡昀翰, “Distributed Astronomy Sequential Pattern Analysis System Using Hadoop Platform with Weighted Suffix Tree,” 國立中央大學, 碩士論文, 2015.