Extracting High Profit Sequential Feature Groups of Products Using High Utility Sequential Pattern Mining

Abstract

Creating a set of product features obtained through mining users’ opinions helps retailers identify the attributes (features or aspects) more accurately and discover the most preferred features of a certain product. High Profit Feature Groups are created by extracting such product feature groups such as ‘{batterylife, camera} of a smartphone,’ which results in higher profit for manufacturers and increased consumer satisfaction. The accuracy of opinion-feature extraction systems can be improved if more complex sequential patterns of customer reviews are included in the user-behavior analysis to obtain relevant feature groups. An existing system referred to in this paper as HPFG19_HU uses High Utility Itemset Mining and Aspect-Based Sentiment Analysis to obtain high profit aspects considering the high utility values, but it does not consider the order of occurrences (sequences) of features formed in customers’ opinion sentences that help distinguish similar users and identify more relevant and related high profit product features. This paper proposes a High Profit Sequential Feature Groups based on the High Utility Sequences (HPSFG_HUS) system, which identifies sequential patterns in features. It combines Opinion Mining with High Utility Sequential Pattern Mining. This approach provides more accurate high feature groups, sales profit, and customer satisfaction, as shown by the retailer’s graphs of extracted High Profit Sequential Feature Groups. Experiments with evaluation results of execution time and evaluation metrics show that this system generates higher revenue than the tested existing systems.

C. I. Ezeife—This research was supported by the Natural Science and Engineering Research Council (NSERC) of Canada under an Operating grant (OGP-0194134) and a University of Windsor grant.