No abstract

When using traditional algorithms to mine the association of hiding information in medical pathological data, there are some problems, such as low recognition rate of association and poor accuracy of mining results. Therefore, structured medical pathology data hiding information association mining algorithm based on optimized convolution neural network is proposed. Firstly, an information feature is optimized based on rough set relative classification information entropy and ant colony algorithm and the optimized feature matrix is obtained. The information in the optimized feature matrix is weighted, and the weighted features of hiding information are obtained. Secondly, the hiding information feature matrix is transmitted to the convolution neural network for learning, and the weight of the connection layer is extracted. The importance of the corresponding area of the weight is confirmed by the distribution of the weight value, and the feature average matrix is obtained. According to the matrix, the feature of hiding information data is enhanced. The hiding information in the structured medical pathology data is generalized by using the Gaussian Bell function, and the hiding information generalization processing result is combined with the adjacent matrix in the convolution neural network to construct the hiding information classification model. Finally, the classification standard is defined, the cooperative association of hiding information group is obtained, and the mining of association between hiding information of structured medical pathological data is completed. The experimental results show that the proposed algorithm has good feature optimization effect, and the information association recognition rate is high, the anti-interference ability and accuracy are better than the current related results, the highest recall rate is 99.24%, which is much higher than the traditional algorithm, which shows that the algorithm is effective. INDEX TERMS Convolutional neural network, deep learning, medical pathology data, hiding information, association mining.

No abstract

No abstract