High Utility Item-set Mining from retail market data stream with various discount strategies using EGUI-tree

High Utility Item-set Mining (HUIM) is the futuristic remodel version of Frequent Item-set Mining (FIM). It discovers customer purchase trends in the retail market. This knowledge is useful to retailers to incorporate various innovative schemes in their businesses to attract the customers such as discounts, cross-marketing, seasonal sale offers…etc. Even though many HUIM algorithms are available to detect profitable patterns, most of them cannot apply to all kinds of retail market data sets due to certain assumptions. The first assumption is that the items always produce a positive profit. Even though purchased items’ overall profit could be positive, few items may have negative profit. Another assumption is they are built for static transactional data. The data is gathered up to the point of time and is used for analysis. It is helpful to make decisions at some intervals like quarterly, half-yearly, yearly. But, to take decisions at any time by analyzing the present sales trend, it is required to process the data stream. This paper presents an innovative idea named Extended Global Utility Item-sets Tree(EGUI-tree) to extract High utility item-sets in the retail market data stream with positive and negative profit items. The sliding window-based technique is applied to the data stream to pick up the very recent data to process. An experimental study on real-world datasets shows that the proposed EGUI-tree algorithm is faster and scalable.