Learning user preference patterns for Top-N recommendations

In this paper, we observe that the user preference styles tend to change regularly following certain patterns. Therefore, we propose a Preference Pattern model to capture the user preference styles and their temporal dynamics, and apply this model to improve the accuracy of the Top-N recommendation. Precisely, a preference pattern is defined as a set of user preference styles sorted in a time order. The basic idea is to model user preference styles and their temporal dynamics by constructing a representative subspace with an Expectation- Maximization (EM)-like algorithm, which works in an iterative fashion by refining the global and the personal preference styles simultaneously. Then, the degree which the recommendations match the active user's preference styles, can be estimated by measuring its reconstruction error from its projection on the representative subspace. The experiment results indicate that the proposed model is robust to the data sparsity problem, and can significantly outperform the state-of-the-art algorithms on the Top-N recommendation in terms of accuracy. © 2012 IEEE.

Single-preference consensus measures based on models of ecological evenness

We investigate the relationship between consensus measures used in different settings depending on how voters or experts express their preferences. We propose some new models for single-preference voting, which we derive from the evenness concept in ecology, and show that some of these can be placed within the framework of existing consensus measures using the discrete distance. Finally, we suggest some generalizations of the single-preference consensus measures allowing the incorporation of more general notions of distance.

A new fuzzy ranking method using fuzzy preference relations

In this paper, a new fuzzy ranking method for both type-1 and interval type-2 fuzzy sets (FSs) using fuzzy preference relations is proposed. The use of fuzzy preference relations to rank FSs with vertices has been introduced, and successfully implemented to undertake fuzzy multiple criteria hierarchical group decision-making problems. The proposed fuzzy ranking method is an extension of the results published in [1], and it is able to rank FSs with and without vertices. Besides that, it is important for a fuzzy ranking method to satisfy six reasonable fuzzy ordering properties as discussed in [6]-[8]. As a result, the capability of the proposed fuzzy ranking method in fulfilling these properties is analyzed and discussed. Issues related to time complexity of the proposed method are also examined.

Preference Networks: probabilistic models for recommendation systems

Abstract
Recommender systems are important to help users select relevant and personalised information over massive amounts of data available. We propose an unified framework called Preference Network (PN) that jointly models various types of domain knowledge for the task of recommendation. The PN is a probabilistic model that systematically combines both content-based filtering and collaborative filtering into a single conditional
Markov random field. Once estimated, it serves as a probabilistic database that supports various useful queries such as rating prediction and top-N recommendation. To handle the challenging problem of learning large networks of users and items, we employ a simple but effective pseudo-likelihood with regularisation. Experiments on the movie rating data demonstrate the merits of the PN.

Preference mining techniques for customer behavior analysis

The thesis has studied a number of critical problems in data mining for customer behavior analysis and has proposed novel techniques for better modeling of the customers’ decision making process, more efficient analysis of their travel behavior, and more effective identification of their emerging preference.

In-store music and aroma influences on shopper behavior and satisfaction

Retail markets are increasingly competitive and retailers continuously look to differentiate their retail offering. One way to differentiate is by providing a pleasant and exciting shopping ambience. This paper experimentally tests the effects of music (volume high or low) and aroma (vanilla scent present/absent) on young fashion shoppers in a real retail setting. Results show that volume of music and the presence of a vanilla aroma both have a significant impact on shoppers' emotions and satisfaction levels. Additional analysis reveals that the arousal induced by music and aroma results in increased pleasure levels, which in turn positively influences shopper behaviors, including time and money spend, approach behavior, and satisfaction with the shopping experience. Direct effects of arousal on behaviors as well as an interaction effect between music and aroma on pleasure and time spent in the store are also present. The paper contributes to the better understanding of shoppers' emotions and shopper behaviors in response to in-store atmospherics and offers retailers practical insights into how to create competitive advantage by customizing the atmosphere in their stores.

Bid chimeras indicate that most BH3-only proteins can directly activate Bak and Bax, and show no preference for Bak versus Bax

The mitochondrial pathway of apoptosis is initiated by Bcl-2 homology region 3 (BH3)-only members of the Bcl-2 protein family. On upregulation or activation, certain BH3-only proteins can directly bind and activate Bak and Bax to induce conformation change, oligomerization and pore formation in mitochondria. BH3-only proteins, with the exception of Bid, are intrinsically disordered and therefore, functional studies often utilize peptides based on just their BH3 domains. However, these reagents do not possess the hydrophobic membrane targeting domains found on the native BH3-only molecule. To generate each BH3-only protein as a recombinant protein that could efficiently target mitochondria, we developed recombinant Bid chimeras in which the BH3 domain was replaced with that of other BH3-only proteins (Bim, Puma, Noxa, Bad, Bmf, Bik and Hrk). The chimeras were stable following purification, and each immunoprecipitated with full-length Bcl-xL according to the specificity reported for the related BH3 peptide. When tested for activation of Bak and Bax in mitochondrial permeabilization assays, Bid chimeras were ~1000-fold more effective than the related BH3 peptides. BH3 sequences from Bid and Bim were the strongest activators, followed by Puma, Hrk, Bmf and Bik, while Bad and Noxa were not activators. Notably, chimeras and peptides showed no apparent preference for activating Bak or Bax. In addition, within the BH3 domain, the h0 position recently found to be important for Bax activation, was important also for Bak activation. Together, our data with full-length proteins indicate that most BH3-only proteins can directly activate both Bak and Bax.

Preference relation-based markov random fields

A preference relation-based Top-N recommendation approach, PrefMRF, is proposed to capture both the second-order and the higher-order interactions among users and items. Traditionally Top-N recommendation was achieved by predicting the item ratings fi rst, and then inferring the item rankings, based on the assumption of availability of explicit feed-backs such as ratings, and the assumption that optimizing the ratings is equivalent to optimizing the item rankings. Nevertheless, both assumptions are not always true in real world applications. The proposed PrefMRF approach drops these assumptions by explicitly exploiting the preference relations, a more practical user feedback. Comparing to related work, the proposed PrefMRF approach has the unique property of modeling both the second-order and the higher-order interactions among users and items. To the best of our knowledge, this is the first time both types of interactions have been captured in preference relation-based method. Experiment results on public datasets demonstrate that both types of interactions have been properly captured, and signifi cantly improved Top-N recommendation performance has been achieved.

Unifying approaches to consensus across different preference representations

© 2015 Elsevier B.V. All rights reserved. Consensus measures can be useful in group decision making problems both to guide users toward more reasonable judgments and to give an overall indication of the support for the final decision. The level of consensus between decision makers can be measured in contexts where preferences over alternatives are expressed either as evaluations or scores, pairwise preferences, and weak orders, however these different representations often call for different approaches to consensus measurements. In this paper, we look at the distance metrics used to construct consensus measures in each of these settings and how consistent these are for preference profiles when they are converted from one representation to another. We develop some methods for consistent approaches across decision making settings and provide an example to help investigate differences between some of the commonly used distances.