In order to reduce the storage pressure of unstructured cloud data and improve the storage capacity of unstructured cloud data,the unstructured cloud data block storage method based on decision tree model is studied.Data cleaning,data selec-tion,data transformation and normalization are used to preprocess unstructured cloud data to reduce the dimension of unstruc-tured cloud data.The method of random feature analysis is adopted to clarify the correlation between the distribution feature quantity of correlation dimensions and the similarity of the unstructured cloud data after preprocessing.Based on this,the fea-tures of unstructured cloud data are extracted by sample expansion and density fusion.The improved decision tree algorithm is used to perform fuzzy classification on the extracted feature set of unstructured cloud data.All kinds of unstructured cloud data are divided into data blocks of the same specification.Through Vandermonde matrix encoding and decoding,unstructured cloud data are stored in blocks on multiple nodes with higher fitness.The experimental results show that the effective calculation ratio of this method reaches 0.8,and it has better storage capacity.The mean compression factor reaches 6.7,which can significant-ly reduce the storage pressure of unstructured cloud data.
关键词
决策树模型/非结构化/云数据/分块存储/预处理/范德蒙矩阵
Key words
decision tree model/unstructured/cloud data/block storage/preprocess/Vandermonde matrix
维普
万方数据