One kind of attribute data based on the multi-segment sbr compression method

The present invention provides a compression algorithm based on SBR segment of multi-attribute data. Mainly comprising: for a plurality of performance data to collect attribute by attribute into the correlation coefficient matrix and a non-signal-yl group signal; a baseband signal represented by the non-return base signal and a regression parameters, if represented by a regression parameters error prediction value and the actual value is greater than the threshold, the segment attributes, and then return said until the error is less than a threshold; yl signals and transmitting only the final regression parameters. The method proposed by the present invention in meeting the requirements of certain errors, performance data can be compressed to be transmitted, saving bandwidth, transmission efficiency is improved performance data.