Web Usage Mining in Noisy and Ambiguous Environments: Exploring the Role of Concept Hierarchies, Compression, and Robust User Profiles

Nasraoui, Olfa; Saka, Esin

doi:10.1007/978-3-540-74951-6_5

Cited by 10 publications

(8 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, relative to the purchase behavior, it carries more noise in the information about preferences. For instance, a customer may search a specific product but eventually not purchase it if she finds out that it is not useful to her [3,35,46]. Hence, we do not consider the customer's search behavior.…”

Section: Objective Function and State Equationmentioning

confidence: 97%

Impact of Recommender System on Competition Between Personalizing and Non-Personalizing Firms

Ghoshal

Kumar

Mookerjee

2015

Journal of Management Information Systems

View full text Add to dashboard Cite

Section: Objective Function and State Equationmentioning

confidence: 97%

Impact of Recommender System on Competition Between Personalizing and Non-Personalizing Firms

Ghoshal

Kumar

Mookerjee

2015

Journal of Management Information Systems

View full text Add to dashboard Cite

“…the presence of outliers. Interestingly, the method can be applied to structured data, such as web-usage logs, to be interpreted in the context of concept taxonomies [32]. Related methods, employed for Web mining tasks, adopt fuzzy membership functions w.r.t.…”

Section: Clustering Proceduresmentioning

confidence: 99%

“…In order to appreciate the type of clustering produced by our algorithm from a more qualitative viewpoint, as an [11,32] example, Fig. 4 reports part of the outcomes of a run on the LUBM ontology, describing the academic domain.…”

Section: Article In Pressmentioning

confidence: 99%

Metric-based stochastic conceptual clustering for ontologies

Fanizzi

d’Amato

Esposito

2009

Information Systems

View full text Add to dashboard Cite

“…Other factors, such as search history, vary over time, and carry less information (relative to actual purchases) about her preferences. For instance, a customer may search for a specific product but not purchase it for a variety of reasons (Alam et al 2013, Nasraoui and Saka 2007, Suryavanshi et al 2005.…”

mentioning

confidence: 99%

Dilemma of Data Sharing Alliance: When Do Competing Personalizing and Non‐Personalizing Firms Share Data

Ghoshal

Kumar

Mookerjee

2020

Production and Operations Management

View full text Add to dashboard Cite

W e analyze a case where two competing firms, a personalizing firm that makes product recommendations and a non-personalizing firm that does not recommend products, compete with each other and may participate in a data sharing alliance (in which the non-personalizing firm shares its data with the personalizing firm). The customer is strategic as she first visits the personalizing firm to find out what product to purchase among the recommended ones and purchases the product eventually from one of the firms. Overall, the customer distributes all her purchases across both the firms-she purchases from the personalizing firm to maintain a profile quality with the firm (to continue to obtain good recommendations) and from the non-personalizing firm to benefit from the lower prices of that firm. We analyze this setup in a two-stage game to find the subgame perfect Nash Equilibrium. In the first stage, firms decide to share data, or not. In the second stage they play a simultaneous move price game. We find that the non-personalizing firm is always willing to share its data with the personalizing firm, but the personalizing firm uses it only when the learning rate influence dominates the profile influence. When the personalizing firm improves its system and both firms share data, its own profit always increases and under certain conditions the profit of the non-personalizing firm also increases. Finally, social surplus may increase as well when firms share data. This result has important implications for the policy-makers controlling information sharing between firms.

show abstract

Web Usage Mining in Noisy and Ambiguous Environments: Exploring the Role of Concept Hierarchies, Compression, and Robust User Profiles

Cited by 10 publications

References 20 publications

Impact of Recommender System on Competition Between Personalizing and Non-Personalizing Firms

Impact of Recommender System on Competition Between Personalizing and Non-Personalizing Firms

Metric-based stochastic conceptual clustering for ontologies

Dilemma of Data Sharing Alliance: When Do Competing Personalizing and Non‐Personalizing Firms Share Data

Contact Info

Product

Resources

About