Finding patterns is one of the auspicious features of data mining. Many researchers invented fabulous methods, algorithms to find frequent patterns as well as regular patterns. More emphasis is required to find regular patterns compared to frequent patterns which bank on traditional support count and confidence frame work. Yet these methods suffer from outliers and require additional means and measures to eliminate them. Hence the FP-Tree and RP tree algorithms are more reliable and effective compared to the earlier algorithms. There were many developments took place in algorithms finding sequential patterns, irregular patterns, and rare patterns and so on. In our studies we observed there was no much emphasis on negative patterns which are very few in nature. Also in real time scenarios there is great scope for the importance of negative patterns for example penguins are birds but do not fly. In this paper we are presenting SW_NPRISM algorithm to find all the negative and positive itemset from a stream of data on sliding window protocol. Data streams are dynamic in nature with numerous variables and complex objects. Sliding window protocol is very much handy in such environment. Our method finds both negative and positive itemset from a stream of data with various regularity thresholds from a window using vertical data format.

Username
Password