天文光谱分类算法在分布式环境下的应用研究

Application Research on Astronomy Spectrum Classification Algorithms under Distributed Environment

以分布式存储的光谱多属性数据集为基础,运用agent技术设计用户代理和挖掘代理,对光谱进行快速并可扩展性的分布式并行处理,减少冗余计算及结点间的通信量,避免了不必要的网络负载和网络竞争的产生,解决了并行分类算法性能的负载均衡问题。

This paper is based on the multi - attribute spectrum dataset and we use user agent and mining agent to fit the distributed mining, reduce the redundancy computing and communications traffic between computing nodes, avoid unnecessary load of network. This paper solves the load balance of parallel classification algorithm to improve the efficiency in mining.