Research On Data Mining Technology And Its Application In Geological Hazard System

China is a country with frequent geological disasters.The common disasters such as mudslides,landslide and ground subsidence and so on,which take great threats and hidden dangers to social development,people's lives and property safety.Therefore,it has prompted many researchers to study geological hazard monitoring and early warning technology.In this paper,the historical data in the geological hazard monitoring system was analyzed,and valuable information was mined by the association rules,so the prediction effect of this system was improved.The core work is to improve the traditional incremental association rule mining FUP algorithm to meet the historical data mining of geological hazard monitoring system.The FUP algorithm has four kinds shortcomings:1)When collecting the frequency of the project collection,the database needs to be scanned frequently;2)When the data is updated,the minimum support threshold is not adjusted with the data updating;3)The positive correlation analysis was not performed for the results of the excavation.In order to improve the effectiveness of this algorithm,the paper discusses the optimization problem of FUP algorithm in the process of acquisition and mining of frequent item sets.The algorithm named IIBTree-FUP(Improved FUP algorithm based on inverted index B-tree)is designed.The main contents of the paper include these aspects:(1)The tree structure is used to store information when counting frequencies on the original data sets.This paper chooses the inverted index tree structure,which based on B+ tree to store the item sets and frequencies that have been obtained.When the data is updated,and new data information is added to the existing tree structure.(2)When the data is updated,the minimum support threshold is predicted and adjusted by Newton interpolation formula,which reduces the time and space complexity of the algorithm to a certain extent,makes the mining association rules more scientific.(3)The positive correlation analysis of the mined association rules is performed to ensure that the selected rules are more scientific and applicable for the geological hazard monitoring system.The improved algorithm not only adaptively adjusts the minimum support threshold,but also reduces the frequent scanning of the database,which greatly improves the scientificity of association rule mining.The improved IIBTree-FUP algorithm is applied to mining association rules in the existing data sets of the geological disaster monitoring system.The experimental results show that the time and space complexity of the IIBTree-FUP algorithm are significantly reduced in the process of frequent item set acquisition,and the potential association rules with positive correlation is obtained.