Analysis of the Algorithm for the Classification and Recognition of the Unbalanced Data Fragments in Large Database
Download as PDF
In the traditional algorithm of data fragment recognition, the influence of fragment's own attributes is often ignored, which leads to the reduction of the accuracy of data fragment recognition. Therefore, we propose a large database based on lrfu strategy and association analysis to identify and classify the unbalanced data segments. Algorithm. Based on the resampling algorithm, the unbalanced data segments are upsampled; the zero filling filter coefficients based on the sampling results and the convolution calculation of the unbalanced data segments based on the filter banks, as well as the reconstruction segments used to obtain the segment feature sequences. The similar partition linear method is used to deal with the highly integrated and unevenly distributed data segments with sequences to achieve segment classification. Transform the objective function, and get the suitable function by combining the extended matrix. The matching function uses lrfu strategy to schedule, correlation analysis to determine the attribute value of the integrated unbalanced data segment, and realize the recognition of the unbalanced data segment. The experimental results show that if the improved method is used to classify and recognize the non-uniform data segments, its suitability and recognition accuracy are high, and it has specific advantages.
Big Database, Imbalance, Data, Fragmentation, Classification, Recognition, Convergence Degree, Attribute Value