A complex panel data learning method based on big data

Abstract

The invention discloses a complex panel data learning method based on big data, which fully utilizes the information increment contained in the newly added time series while controlling the time series data capacity, thereby controlling the calculation cost and computer storage capacity; making full use of the cross-sectional data While containing dimension information, control the capacity of the supporting dimension, so as to control the computing cost and storage capacity within the permitted range of computing resources; under the condition of limited panel data capacity, the generalization performance of learning is guaranteed, and the generalization performance of the proposed algorithm is given. Research on the theory and method of complex panel data white noise filtering, to ensure the robustness of the learning model, using dynamic vector fusion technology, the calculation performance is robust, time series and cross-sectional high-dimensional complex panel data problems, using a linear learning model , realizing online kernel learning whose bidirectional data capacity is simultaneously controlled.

Description

technical field [0001] The invention relates to a panel data learning method, in particular to a panel data learning method based on big data. Background technique [0002] The research on panel data is mostly based on econometric modeling theory and application, generally there are assumptions about the data model, and it is a batch data learning mode. However, with the complexity of social and economic relations, panel data presents many new features, mainly in terms of large scale of data, intricate relationships, serious data pollution and missing data. In today's big data environment, the original The model assumptions of , do not necessarily hold in the context of big data. At the same time, the limitation of computer storage capacity also restricts the normal use of the original panel data model. SUMMARY OF THE INVENTION [0003] In order to solve the problems of serious pollution of existing panel data and missing data, the purpose of the present invention is to ...

AI Technical Summary

Problems solved by technology

[0003] In order to solve the problems of serious pollution and missing data in the existing panel data, the purpose of the present invention is to provide a complex panel data learning method based on big data, seek a complex panel data processing mode that reduces computing costs and saves computer storage resources, and uses Solve the problems raised in the above-mentioned background technology