Scalable Vertical Mining for Big Data Analytics of Frequent Itemsets

Advances in technology and the increasing growth of popularity on Internet of Things (IoT) for many applications have produced huge volume of data at a high velocity. These valuable big data can be of a wide variety or different veracity. Embedded in these big data are useful information and valuable knowledge. This leads to data science, which aims to apply big data analytics to mine implicit, previously unknown and potentially useful information from big data. As a popular data analytic task, frequent itemset mining discovers knowledge about sets of frequently co-occurring items in the big data. Such a task has drawn attention in both academia and industry partially due to its practicality in various real-life applications. Existing mining approaches mostly use serial, distributed or parallel algorithms to mine the data horizontally (i.e., on a transaction basis). In this paper, we present an alternative big data analytic approach. Specifically, our scalable algorithm uses the MapReduce programming model that runs in a Spark environment to mine the data vertically (i.e., on an item basis). Evaluation results show the effectiveness of our algorithm in big data analytics of frequent itemsets.