Process mining for recommender strategies support in news media

Abstract

The strategic transition of media organizations to personalized information delivery has urged the need for richer methods to analyze the customers. Though useful in supporting the creation of recommender strategies, the current data mining techniques create complex models requiring often an understanding of techniques in order to interpret the results. This situation together with the recommender technologies deluge and the particularities of the news industry pose challenges to the news organization in making decisions about the most suitable strategy. Therefore, we propose process mining as a high-level, end-toend solution to provide insights into the consumers' behavior and content dynamics. Specifically, we explore if it allows news organizations to analyze independently and effectively their data in order to support them in defining recommender strategies. The solution was implemented in a case study with the third largest news provider in Norway and yielded preliminary positive results. To our knowledge, this is the first attempt to apply a process mining methodology and adapt the techniques to support media industry with the recommender strategies.

Recommending Personalized News in Short User Sessions

“…Common approaches discover readers' interests from click behavior and articles' categories [23]. In some cases, the categories of the articles are already de ned in advance and represented as contextual meta-data [6,9,10]. In other cases, categories are discovered automatically and represented either as more granular vocabularies associated to the categories [2,4,25], or keywords de ning general, well-known topics such as sport or politics [24,26,37].…”

“…us, in practice, publishers observe relatively short sessions with fewer than ten clicks on average [9,10,27].…”

Recommending news in traditional media companies

“…In addition to users' long-term profiles, some news recommender systems address this with the introduction of short-term profiles for users' current preferences. Experiments also show that many readers follow certain topical paths when they browse the news, which can be explored used process mining techniques Epure et al 2016). An analysis of Dagbladet's user logs show that their readers are almost four times more likely to move from political to international news than in the other direction.…”

“…Experiments also show that many readers follow certain topical paths when they browse the news, which can be explored used process mining techniques (e.g. Epure et al 2016). An analysis of Dagbladet's user logs show that their readers are almost four times more likely to move from political to international news than in the other direction.…”

Modeling the Dynamics of Online News Reading Interests

“…In some cases, news category taxonomies are created in advance. en, representative categories are selected for each article and included in the article's meta-data [3,5]. In other cases, news categories are automatically identi ed from the article's content as either granular vocabularies [1] or broad news topics such as tourism or local news [6,7].…”