A novel classification-based shilling attack detection approach for multi-criteria recommender systems

Recommender systems are emerging techniques guiding individuals with provided referrals by considering their past rating behaviors. By collecting multi-criteria preferences concentrating on distinguishing perspectives of the items, a new extension of traditional recommenders, multi-criteria recommender systems reveal how much a user likes an item and why user likes it; thus, they can improve predictive accuracy. However, these systems might be more vulnerable to malicious attacks than traditional ones, as they expose multiple dimensions of user opinions on items. Attackers might try to inject fake profiles into these systems to skew the recommendation results in favor of some particular items or to bring the system into discredit. Although several methods exist to defend systems against such attacks for traditional recommenders, achieving robust systems by capturing shill profiles remains elusive for multi-criteria rating-based ones. Therefore, in this study, we first consider a prominent and novel attack type, that is, the power-item attack model, and introduce its four distinct variants adapted for multi-criteria data collections. Then, we propose a classification method detecting shill profiles based on various generic and model-based user attributes, most of which are new features usually related to item popularity and distribution of rating values. The experiments conducted on three benchmark datasets conclude that the proposed method successfully detects attack profiles from genuine users even with a small selected size and attack size. The empirical outcomes also demonstrate that item popularity and user characteristics based on their rating profiles are highly beneficial features in capturing shilling attack profiles.     

Determination of optimal product styles by ordinal logistic regression versus conjoint analysis for kitchen faucets

In this study, a two-stage integrated approach is proposed and implemented to explore user perceptions about kitchen faucet styles and to find optimal levels of design parameters related to product appearance. At the first stage, a group of representative users have been asked to judge 38 systematically selected different faucet designs by using a semantic differential (SD) scale for 11 image (kansei) words about their visual perceptions. Then the relations between overall preference and kansei word scores of users are investigated by Ordinal Logistic Regression (OLOGREG). The model obtained helps understand how customers describe highly preferred designs. It can also be used to confirm if a product design will be highly preferred. At the second stage, the relations between overall preference scores and design parameter levels related to the product appearance are modeled using again OLOGREG, and then the best design parameter levels for visual perceptions are found by maximizing the overall preference scores. The results are confirmed and discussed. They are also compared to those of a commonly used approach in the literature, Conjoint Analysis (CA). This comparison has showed that the OLOGREG approach is superior to traditional CA. Finally, it is discussed how the product style design optimization approach presented and demonstrated in this study for the case of some kitchen faucets can be used in general for other products.Relevance to industry: Developing a user-centered product is an important policy of an enterprise in today's highly competitive marketplace. To reach this objective, a systematic method is proposed by integrating several techniques. This method will be useful to any industry that designs and produces consumer products. These products more fit the consumer needs and the competitiveness of them are improved.     

Multi-period lot-sizing with supplier selection using achievement scalarizing functions

In this paper, an integration of Analytic Network Process (ANP) and achievement scalarizing functions is proposed to choose the best suppliers and define the optimum quantities among the selected suppliers by considering tangible–intangible criteria and time horizon. To reflect the decision maker’s (DM’s) preferences more accurate, an additive achievement function is defined consist of several components. In this additive function while unwanted deviations from periodic budget and aggregate quality goals are balanced by Minmax Goal Programming (MGP), and unwanted deviations from total cost, total value of purchasing (TVP) and aggregate quality are minimized by Achimedean Goal Programming (AGP) to provide more acceptable solutions. This multi-period model enables us to reflect DM’s preferences more flexible than the other traditional models that use only one type of achievement function. The sensitivity analysis was also performed for different levels of periodic demands. It is also possible to enlarge the sensitivity analyses for other parameters such as different levels of capacity, and different weights of components.     

Microtubule-Targeting 7-Deazahypoxanthines Derived from Marine Alkaloid Rigidins: Exploration of the N3 and N9 Positions and Interaction with Multidrug-Resistance Proteins

Our laboratories have been investigating synthetic analogues of marine alkaloid rigidins that possess promising anticancer activities. These analogues, based on the 7-deazahypoxanthine skeleton, are available in one- or two-step synthetic sequences and exert cytotoxicity by disrupting microtubule dynamics in cancer cells. In the present work we extended the available structure–activity relationship (SAR) data to N3- and N9-substituted derivatives. Although N3 substitution results in loss of activity, the N9-substituted compounds retain nanomolar antiproliferative activities and the anti-tubulin mode of action of the original unsubstituted compounds. Furthermore, our results also demonstrate that multidrug-resistance (MDR) proteins do not confer resistance to both N9-unsubstituted and -substituted compounds. It was found that sublines overexpressing ABCG2, ABCC1, and ABCB1 proteins are as responsive to the rigidin analogues as their parental cell lines. Thus, the study reported herein provides further impetus to investigate the rigidin-inspired 7-deazahypoxanthines as promising anticancer agents.     

Properties of the maximum likelihood and pisarenko spectral estimates

The performance of the maximum likelihood spectral estimate is compared vis-à-vis the classical Bartlett spectrum. Asymptotic behavior and special cases such as single tone harmonic analysis are considered. Also, the Pisarenko method of harmonic retrieval is analyzed for the case of a single tone imbedded in noise. The statistical properties of the eigenvector corresponding to the minimum eigenvalue as well as that of the estimated frequency value are derived. The dependence of the reliability of the estimate upon the data size and the tone frequency location is illustrated.     

Shear amplification of orientation in Nylon 11/clay nanocomposites during melt casting and resulting electrical properties

In this study, the effect of organically modified clay on the orientation enhancement in Nylon 11 in melt casting was investigated. Nylon 11 was mixed with 1 and 3 wt% Cloisite 20A using twin screw extrusion and they were cast into films with varying take-up speeds. The addition of clay in Nylon 11 helped increase orientation levels substantially in melt cast films, both as a function of clay concentration as well as take-up speeds. This was primarily due to shear amplification effect caused by the movement of adjacent clay nanoparticles due to the shear flow gradient within the die. At low clay concentrations, the sub-T_m stretchability, and electrical breakdown strength improve as the presence of clay reduces inter/intrachain hydrogen bonding. At higher clay concentrations, both orientation and electrical breakdown levels decrease. The latter is primarily caused by increased percolation path of charge carriers. Nevertheless, clay nanoplatelets were very effective in their role as melt processing aids, as they enhance orientation levels of Nylon 11 thin films by shear amplification effect where they increase local chain orientation of chains trapped between clay platelets while their orientation relaxation is suppressed.     

The Use of Digital Image Processing Method to Estimate the Foam Characteristics of Polyurethane

Four air-bubbled polyurethane (PU) foams with different polyol:PMDI wt.% are produced, respectively. The chemical reaction mechanisms of polyurethane and bubble formation are proposed by performing standard Gibbs free energy calculations using the DFT M06-2X/6-31+G(d,p) method. The local minima, transition states, and intermediates in reaction mechanisms are detected. It is concluded that both reactions are exothermic. Then, raw images of the produced PU foams are taken with a 13 MP mobile phone camera, which can be considered inexpensive, and the mean radii of the pores are calculated by an image processing based method (IPBM) on a standard desktop computer with an i5 processor. It is determined that there is a close relationship between the calculated mean radius and instrumentally measured thermal conductivity coefficient of the foams. However, the thermal conductivity coefficients are independent of the calculated number and percentage of the pores. The mean radii of the samples calculated by the proposed IPBM are close to that of the SEM, with acceptable relative errors of less than 10%. Finally, it is concluded that IPBM, which is a more cost-effective, cleaner, and faster method than SEM, might replace SEM in the air bubble analysis of PU foams.