Multi-algorithm-integrated big data analysis middleware platform

Abstract

The invention discloses a multi-algorithm-integrated big data analysis middleware platform, and belongs to the technical field of software development. The platform comprises an algorithm realization layer, an algorithm support layer and an algorithm, storage layer, wherein the algorithm support layer is interacted with the algorithm realization layer through a standard interface; the algorithm support layer is used for combining algorithms to form demand algorithms required by new businesses; the algorithm storage layer is used for carrying out server side storage on various algorithms mounted on the platform. According to the platform, functions such as algorithm mounting, algorithm running and algorithm evaluation are developed for the big data analysis through a series of multi-algorithm-integrated developments. For the superior, complicated algorithms are shielded, and for the inferior, specific businesses are shielded. Through combining various algorithms mounted on the platform, solutions required by the businesses are rapidly formed, various algorithms can be operated, and the algorithm operation effect can be evaluated.

Description

technical field [0001] The invention relates to the field of software architecture and technology, in particular to a big data analysis middleware platform integrating multiple algorithms. Background technique [0002] In today's society, various industries have widely used computer software to realize their specific business needs. Good program design is the prerequisite for realizing business requirements, and algorithm is the soul of the program. In practical applications, different programming languages ​​can be used to implement the algorithm. However, in many cases, the algorithm has problems of poor portability and scalability because it meets business needs too much. Even by optimizing parameters, refactoring code, and choosing a more reasonable data structure, it is difficult to meet today's geometrically growing data volume and various actual business needs of various industries. [0003] In the era of big data, for various industries, if we develop special big d...

AI Technical Summary

Problems solved by technology

However, this paper mainly studies the reasonable collocation of algorithms and data in the process of information fusion, not the platform that supports algorithm development and the isolation of algorithms and applications.

[0006] In view of the above background and practical problems faced, the present invention proposes a big data analysis middleware design integrating multiple algorithms, using the middleware to shield the underlying differences, and being able to run on various operating system platforms, etc., aiming to solve the problem of algorithms and actual business The problems of high coupling, difficult portability, and low expansion

Images