Data recognition method and device

Abstract

The invention discloses a data recognition method and device, and relates to the technology of data processing. A label data set including label data samples is divided into a plurality of data subsets according to the difference of the label data samples, so that all recognizers conduct training according to the data subsets respectively, the difference of all the recognizers is guaranteed, hence, when data recognition is carried out on data to be recognized, given recognition results of the recognizers after training are obtained, then, a final recognition result of the data to be recognized is determined according to all the recognition results, and thus the accuracy of big data recognition is improved.

Description

Technical field [0001] The present invention relates to data processing technology, in particular to a data identification method and device. Background technique [0002] At present, the data generation speed in the real and virtual world is increasing. Automatic identification of data will facilitate users to find and use the data. Therefore, when many applications or systems obtain new data, they need to pass on existing data. Recognition method, data recognition of the obtained data. [0003] The current data recognition method is mainly: first select the corresponding training data from the labeled data to be learned by the recognizer, and when new data is obtained, the learned recognizer can be used to recognize the data. [0004] When learning the recognizer, the technologies most relevant to this patent include large-scale machine learning and integrated learning. The two learning methods are specifically explained below: [0005] Large-scale machine learning refers to the th...

AI Technical Summary

Problems solved by technology

[0009] Integrated learning methods have strong scalability for machine learning on large-scale data, but if they are directly applied to large-scale data learning problems, the final data mining accuracy will not be very high due to the inconspicuous differences , can not achieve the desired effect

Although large-scale data learning is achieved, it cannot fully reflect the advantages of large-scale data learning

[0010] At present, the ensemble learning method can be applied to large-scale learning through resampling technology and subset division. However, different recognizers can be sampled in big data through different sampling techniques to achieve the trained The difference of the recognizer, but because only part of the training data is sampled and used for the construction of the recognizer, the knowledge contained in the labeled samples is not used for the training of the recognizer

However, directly dividing the entire labeled data set into multiple subsets, this method of random labeling data set splitting cannot guarantee the difference between different sub-labeled sets, so that the difference between multiple recognizers cannot be guaranteed.

Images