Using Glocal Event Alignment for Comparing Sequences of Significantly Different Lengths
This work takes place in the context of conversion rate optimization by enhancing the user experience during navigation on e-commerce web sites. The requirement is to be able to segment visitors into meaningful clusters, which can then be targeted with sp
- PDF / 637,036 Bytes
- 15 Pages / 439.37 x 666.142 pts Page_size
- 5 Downloads / 152 Views
act. This work takes place in the context of conversion rate optimization by enhancing the user experience during navigation on e-commerce web sites. The requirement is to be able to segment visitors into meaningful clusters, which can then be targeted with specific call-to-actions, in order to increase the web site turnover. This paper presents an original approach, which equally combines global- and local-alignment techniques (Needleman-Wunsch and Smith-Waterman) in order to automatically segment visitors according to the sequence of visited pages. Experimental results on synthetic datasets show that our approach out-performs other typically used alignment metrics, such as hybrid approaches or Dynamic Time Warping. Keywords: Web mining
1
· Sequential pattern mining · Clustering
Introduction
Conversion rate optimization is considered as one of the most promising approaches for improving the turnover of e-commerce web sites. A lot of researches have already focused on understanding web browsing event-patterns, in order to improve the online content delivery. Clustering visitors into meaningful segments, associated to targeted call-to-actions and related item recommendation, is one of the techniques typically used for cross- and up-selling. As clustering aims at organizing similar items into the same group with no prior knowledge of item class, it is seen as an approach of unsupervised learning. In web usage mining context, the similarity of page visits and their order in a session is one of the relevant information to cluster. For example, cluster analysis helps to reach people who are interested in some specific kind of goods or services so that the owner can recommend to such groups other related things, or offer them some discounts. The clustering result can also be applied to advertising placement organization on web sites, based on page visiting frequency in each cluster. In this paper, we present our work for computing the similarity between event-sequences of significantly different lengths. Our proposal is based on c Springer International Publishing Switzerland 2016 P. Perner (Ed.): MLDM 2016, LNAI 9729, pp. 58–72, 2016. DOI: 10.1007/978-3-319-41920-6 5
Using Glocal Event Alignment for Comparing Sequences
59
a new way of equally combining global- and local-alignment techniques (i.e. Needleman-Wunsch [21] and Smith-Waterman [29]). The originality of our measure is to take into account the length of longest sequence in the pair of compared sequences. Thus, regardless of the difference in sequence lengths, the result provided by our metric is accurate and can be used to perform clustering. Experimental results show that our approach outperforms other typically used similarity measures, such as hybrid approaches or Dynamic Time Warping (DTW), in the context of event-sequences of different lengths. This paper is divided into five sections with the following structure: Section 2 explains the proposed method. Section 3 describes experimental results. The discussion of these results is in Section 4. Section 5 pres
Data Loading...
