Research On Fast Frequent Itemsets Mining Algorithm And Their Applications

With the development of information technology,data grows exponentially,traditional data mining algorithm is inefficient and realizes application difficultly,when it processes and applies to large amounts of data.Data generation,collection,storage and analysis are supported by big data technology,so how to effectively incorporating big data technology into traditional data mining algorithm has become a research trend.The existing frequent itemset mining algorithm has a long runtime or memory overflow when it process in big data.Incorporating big data technology into them effectively,which is helpful to realize big data industry application.On the one hand,big data technology makes data analysis and data processing speed up,which improves the efficiency of algorithm.On the other hand,through analyzing the implicit value of data and applying,it will be convenience to producing and living.This paper based on frequent itemset mining and big data technology explores improved algorithm with higher operating efficiency,finally applying it to reality scene.Specifically,the contribution of this paper are as follows:(1)Improved Eclat algorithmBy expounding Eclat algorithm,optimizing algorithm in two aspects: optimized candidate set and pruning which are the properties of frequent itemset.We screen candidate set and exclude non-candidate set to reducing the times of intersection operations,we propose Eclat' algorithm.The performance of the improved algorithm is verified well by experiments on public data sets.(2)Parallel implementation of improved algorithmIn order to analyze data and utilize data effectively,parallel algorithm is proposed based on Map Reduce experimented on the Hadoop cluster.It is shown that parallel algorithm has a good performance for large data than original algorithm.(3)ApplicationConducting applied research on data of mobile phone users,discussing the relationship between different attributes.It is used to recommend APP to users reasonably,finally making a further exploration on the application of smart city.