Maximal pattern mining method for uncertain data based on depth-first

Abstract

The invention relates to a maximal pattern mining method for uncertain data based on depth-first. The maximal pattern mining method comprises three major technical parts of uncertain data processing, frequent item set judgment and the maximal pattern mining method. The uncertain data processing refers to converting an uncertain data horizontal format of which the main key is an affair ID into an uncertain data vertical format of which the main key is an item ID by virtue of data vertical format conversion. The frequent item set judgment refers to the process of calculating whether the support degree of an item set is greater than or equal to a given support degree threshold and whether the confidence degree of the item set is greater than or equal to a given confidence degree threshold. The maximal pattern mining method is the process of mining the maximal frequent item set, and in the mining process, the converted vertical-format data is taken as the input, and all the uncertain data maximal pattern frequent item sets are mined out according to the given support degree and confidence degree thresholds. The maximal pattern mining method for the uncertain data based on depth-first is capable of effectively obtaining the value information in the uncertain data, and also has high mining efficiency.

Description

technical field [0001] The invention relates to an algorithm in the field of computer application technology, in particular to a depth-first-based mining method for the largest mode of uncertain data. Background technique [0002] With the rapid development of science and technology, technologies such as sensor networks, radio frequency identification, and privacy protection have been extensively researched and applied, followed by such a type of data that does not exist in the form of a single data point, but rather Appears on multiple data points with a certain probability. This is essentially different from the data in traditional databases, and people call it uncertain data. [0003] The mining of frequent itemsets is a basic and core issue in the field of data mining. The representative methods for mining frequent itemsets, such as Max-Miner, Mafia, and Genmax, have been proposed so far. Depth-Project, local maximum frequent itemsets for superset testing, unsupported ...

AI Technical Summary

Problems solved by technology

But the defect of this existing technology is: it can only be applied to deterministic data structure, it mainly judges whether it is frequent by whether the number of occurrences of an item is greater than a given threshold, but when an item occurs with a certain probability, this method does not then apply

In addition, this method uses a recursive method to carry out deep mining, and can only go back one step at a time, and there will be many unnecessary branch judgments, thereby reducing the mining efficiency

Images