A Multi-Source Data Aggregation and Multidimensional Analysis Model for Big Data

• Main Authors: Liu Pan, Chen Lin
• Format: Article
• Language: English
• Published: EDP Sciences 2017-01-01
• Series: ITM Web of Conferences
• Online Access: https://doi.org/10.1051/itmconf/20171205009

With the rise of Internet applications such as search engines, social networks, and e-commerce, the amount of data in the Internet is rapidly expanding. There are a lot of data generated every moment, and the global information is also increasing. Therefore, the big data is driven by the Internet industry, and it is also a subversive technological innovation compared to the cloud computing and Internet of things. How to carry on the fast retrieval in the massive and different types of data? How to discover potential associations between different data? How to mining the potential value of the data? And how to create a multidimensional view of the data? These urgent problems need to be solved. In this paper, a multi-source data aggregation and multidimensional analysis model for big data (DAM_AM) is proposed. The model adopts the hierarchical structure and introduces data aggregation mechanism, multi-source processing mechanism, object and association mapping mechanism, and "walk" mechanism. Using these mechanisms, multi-source data is normalized to a coherent and consistent representation pattern. And then the fields that represent a class of entity in the representation pattern are aggregated into a set of fields. By mapping different field sets into different objects and associations and combining with the time dimension and space dimension, we can build a multifaceted visual model. Through the concrete case analysis and verification, it indicates that the DAM_AM model can analyze the data from multidimensional and multi-level, and shows the potential correlation between different data. The model not only has high computational efficiency and has high scalability, but also shows the analysis results clearly and intuitively.