Frequent itemset mining discovers implicit, previously unknown and potentially useful knowledge—in the form of frequent itemsets—from data. For example, discovery of frequently purchased merchandise products reveals customer purchase patterns, which help store managers about their business strategies and promotional tactics. These, in turn, help increase profits of the stores. As another example, discovery of popular collections of courses reveals course popularity and trends of some subject matters. These, in turn, assist university administrators schedule courses and their corresponding exams to avoid conflict or exam hardship, as well as planning of the calendar.
As we are living in the era of big data, many applications and services generate high volumes of a wide variety of highly valuable data at a high velocity. These data can be of a wide range of veracity. Consequently, having scalable frequent itemset mining service is important to both the data mining experts and non-experts.
Over the past two decades, numerous frequent itemset mining algorithms have been proposed. Many of them require some degrees of data mining knowledge and expertise, which may be inaccessible by layman.
In this paper, we propose a tool with an intention to provide scalable frequent itemset mining-as-a-service (FIMaaS) on cloud for non-expert data miners.