Data processing method and device and device for data processing

Abstract

The embodiment of the invention provides a data processing method and device and a device for data processing. The method comprises the following steps: collecting a feature value set of each data party in m data parties, wherein feature values in the feature value sets are ciphertexts; combining the feature value sets to obtain a combined set; according to the combined set, a characteristic value union set corresponding to each data party in the m data parties is determined, and the characteristic value union set of the i-th data party is obtained by replacing characteristic values of non-i-th data parties in the combined set with placeholders; and returning the corresponding characteristic value union set to each data party in the m data parties, so that each data party performs one-hot coding on the received characteristic value union set. According to the embodiment of the invention, the characteristic value of each data party can be ensured not to be exposed to other data parties, and the data privacy safety of united one-hot coding can be ensured.

Description

technical field [0001] The present invention relates to the field of computer technology, in particular to a data processing method, device and device for data processing. Background technique [0002] One-hot encoding is One-Hot encoding, also known as one-bit effective encoding. The method is to use N-bit state registers to encode N states. Each state has an independent register bit, and at any time, only one of them bit is valid. [0003] For feature data with n values, after one-hot encoding, an n-dimensional 0-1 vector is obtained, and at any time, there will only be one 1 in this vector. For example, for the feature data representing color, there are three values ​​as follows: red=1, green=2, blue=3, after one-hot encoding, the following 0-1 vector can be obtained: red=(0,0, 1), green=(0,1,0), blue=(1,0,0). One-hot encoding converts the values ​​of discrete features into sparse data, which solves the problem that classifiers cannot handle attribute data well, and al...

AI Technical Summary

Problems solved by technology

[0004] However, when performing joint one-hot encoding on multi-party data, such as joint one-hot encoding of data held by party A and data held by party B, it will result in party A's data being exposed to party B, and party B's data Exposure to Party A may result in leakage of private data

Images