Applied Research And Implementation Of Incremental Association Rule Algorithm In Mining Mobile Phone Virus

Mobile phone virus spreads more and more widely while the variety of the virus emerges in endlessly, which making people pay more attention to the safe use of mobile phones. The mobile phone virus prevention technology has also been constantly upgraded. There is an urgent need to resist the invasion of the virus in order to avoid disclosing mobile phone privacy and causing economic losses for users. So, how to quickly excavate the mobile phone virus becomes more and more important. For the anti-virus companies, it is an important task to quickly excavate the mobile phone virus. Simultaneously, the companies need constantly update the virus database in order to guard against new viruses.At present, various data mining techniques have been used to mining the mobile phone virus. Among that, association rule mining is a hot research in recent years. Apriori algorithm is a classical algorithm for mining association rules, by finding frequent itemsets and generating association rules. Finally, we match sample data based association rules to detect the virus. The Apriori algorithm is an algorithm that uses the support-confidence model, but it has some shortcomings, including repeatedly scanning the database, generating a lot of useless rules, and cannot handle incremental data sets. Piatetsky-Shapiro proposed PS model, which improve the validity of rules generated by the support-confidence model. It greatly improves quality and interestingness of the rules. It uses interestingness threshold to filter rules in order to reduce the number of association rules, which are useless, inappropriate. PS model can generate interesting association rules, but did not make full use of results. In the incremental mining process it does not use the itemsets and frequent itemsets generated by the previous process, so the incremental mining will need re-excavation historical data. So it is not strong scalability in the practical application.In order to improve the shortcomings of the classical Apriori algorithm and remedy the flaw of the PS model, we can use the incremental mining algorithm. Its features are merging the concept of accuracy and considering the objective factors like as data inherent relation and the subjective factors like as user’s knowledge and social experience. After adding the subjective factors, it further improves the interestingness of the rules and reduces lots of unwanted or rubbish rules. List structure used by the incremental mining algorithms is easy to understand and implement. The k-itemsets can be classified into different lists and stored into the database simultaneously. In addition, during the increment mining process, it designs special list structure. It can make full use of the results which have been excavated in the priori mining process with special design of list structure. Furthermore, when generating association rules, the algorithm greatly improves the processing efficiency with only dealing with part itemsets to avoid repeatedly generating rules from some frequent itemsets. And the algorithm can update association rules base and delete those unwanted or rubbish rules so that rules in this base are interesting and effective.