基于深度学习和时空关联的大数据分析系统研究

Research on Big Data Analysis System based on Deep Learning and Spatio-Temporal Correlation

基于深度学习、时空关联、大数据技术构建非结构化和结构化数据关联分析的大数据分析系统,提升数据价值,适应多场景应用。智能采集设备采集视频、图像等非结构化数据,同时可以采集经纬度、操作信息、电子设备ID等结构化数据。此外,提出一种基于深度学习进行非结构化数据的实体识别和实体归一,基于同一时空和发掘数据间的关联关系,通过大数据分析形成聚合档案,提供多维度关联分析能力的方法。最后,介绍了基于深度学习和时空关联的大数据分析系统的架构设计,并阐述了系统的关键算法和实现。

Based on deep learning,spatio-temporal correlation and big data technology,building a big data analysis system for unstructured and structured data correlation analysis can enhance the value of data and adapt to multi-scenario applications.At the same time,Intelligent acquisition equipments collect unstructured data such as videos and images,and can collect structured data such as latitude and longitude,operating information,and electronic device IDs.This paper proposes a method for entity recognition and entity normalization of unstructured data based on deep learning,mining the association relationship between data based on the same time and space,forming aggregate files through big data analysis,and providing multi-dimensional association analysis ability.Finally,it introduces the architecture design of big data analysis system based on deep learning and spatio-temporal correlation,and expounds the key algorithms and implementation of the system.