Item Profiles Research Articles

Exploiting interactions of review text, hidden user communities and item groups, and time for collaborative filtering

Rich side information concerning users and items are valuable for collaborative filtering (CF) algorithms for recommendation. For example, rating score is often associated with a piece of review text, which is capable of providing valuable information to reveal the reasons why a user gives a certain rating. Moreover, the underlying community and group relationship buried in users and items are potentially useful for CF. In this paper, we develop a new model to tackle the CF problem which predicts user's ratings on previously unrated items by effectively exploiting interactions among review texts as well as the hidden user community and item group information. We call this model CMR (co-clustering collaborative filtering model with review text). Specifically, we employ the co-clustering technique to model the user community and item group, and each community---group pair corresponds to a co-cluster, which is characterized by a rating distribution in exponential family and a topic distribution. We have conducted extensive experiments on 22 real-world datasets, and our proposed model CMR outperforms the state-of-the-art latent factor models. Furthermore, both the user's preference and item profile are drifting over time. Dynamic modeling the temporal changes in user's preference and item profiles are desirable for improving a recommendation system. We extend CMR and propose an enhanced model called TCMR to consider time information and exploit the temporal interactions among review texts and co-clusters of user communities and item groups. In this TCMR model, each community---group co-cluster is characterized by an additional beta distribution for time modeling. To evaluate our TCMR model, we have conducted another set of experiments on 22 larger datasets with wider time span. Our proposed model TCMR performs better than CMR and the standard time-aware recommendation model on the rating score prediction tasks. We also investigate the temporal effect on the user---item co-clusters.

N-dimensional Markov random field prior for cold-start recommendation

A recommender system is a commonly used technique to improve user experience in e-commerce applications. One of the popular recommender methods is Matrix Factorization (MF) that learns the latent profile of both users and items. However, if the historical ratings are not available, the latent profile will draw from a zero-mean Gaussian prior, resulting in uninformative recommendations. To deal with this issue, we propose using an n-dimensional Markov random field as the prior of matrix factorization (called mrf-MF). In the Markov random field, the attribute (such as age, occupation of users and genre, release year of items) is considered as the site and the latent profile, the random variable. Through the prior, new users or items will be recommended according to its neighbors. The proposed model is suitable for three types of cold-start recommendation: (1) recommend new items to existing users; (2) recommend new users for existing items; (3) recommend new items to new users. The proposed model is assessed on two movie datasets, Movielens 100K and Movielens 1M. Experimental results show that it can effectively solve each of the three cold-start problems and outperforms several matrix factorization based methods.

Anxiety and oppositional behavior profiles among youth with selective mutism

Selective mutism (SM) is a debilitating condition in which a child does not speak in social situations where speech is expected. The clinical conceptualization of SM has been debated historically, with evidence pointing partly to anxious and oppositional behavior profiles. Behavioral characteristics were examined in a clinical sample of 57 youth formally diagnosed with selective mutism. Parents rated children across internalizing and externalizing behaviors on the Child Behavior Checklist. Eighteen highly rated items were subjected to exploratory and then confirmatory factor analysis. Anxiety and oppositional behavior factors were derived. The anxious behavior profile was associated with social anxiety disorder symptoms, social problems, and aggressive behaviors but not oppositional defiant disorder symptoms. The oppositional behavior profile was associated with aggressive behaviors, oppositional defiant disorder symptoms, social problems, and inversely to social anxiety disorder symptoms. Results are consistent with emerging research regarding subgroups of children with SM. Behavior profiles are discussed as well with respect to assessment and treatment implications.Learning outcomes: Readers will learn about the nature of children with selective mutism as well as behaviors that differentiate anxious and oppositional behavior profiles. Items that comprise anxious and oppositional behavior profiles are presented. These item profiles may have ramifications for assessment and treatment.

An NCME Instructional Module on Latent DIF Analysis Using Mixture Item Response Models

The purpose of this ITEMS module is to provide an introduction to differential item functioning (DIF) analysis using mixture item response models. The mixture item response models for DIF analysis involve comparing item profiles across latent groups, instead of manifest groups. First, an overview of DIF analysis based on latent groups, called latent DIF analysis, is provided and its applications in the literature are surveyed. Then, the methodological issues pertaining to latent DIF analysis are described, including mixture item response models, parameter estimation, and latent DIF detection methods. Finally, recommended steps for latent DIF analysis are illustrated using empirical data.

National Institutes of Health Stroke Scale Item Profiles as Predictor of Patient Outcome: External Validation on Safe Implementation of Thrombolysis in Stroke-Monitoring Study Data.

National Institutes of Health Stroke Scale (NIHSS) item profiles that were recently proposed and validated may prove useful for clinical prognostication and research studies. We aimed to validate the NIHSS item profiles in hyper-acute stroke patients who received thrombolysis treatment (tissue-type plasminogen activator). We applied the latent class analysis probabilities of the profile membership generated from the derivation study onto NIHSS data from the Safe Implementation of Thrombolysis in Stroke-Monitoring Study (SITS-MOST). We separately considered NIHSS data collected within 3 hours and at ≈24 hours after stroke onset to obtain 2 sets of symptom groupings. The discrimination and calibration of both sets of symptom profiles were assessed from their association with outcomes. The outcome measures included modified Rankin Scale (mRS; using full distribution and dichotomized, mRS 0-1 or back to baseline) at day 90 and mortality by 90 days. We obtained data for 6843 patients. Ordinal analysis of mRS showed odds of better outcome across the profiles, for each set of symptom profiles, adjusted for age, sex, and prestroke mRS. Dichotomized outcomes mirrored the ordinal findings. There were significant differences in prognostic discrimination ability for the dichotomized outcome measures between the 2 sets of symptom profiles, with the latter set (ie, 24-hour symptom profiles) performing better. The NIHSS item profiles are individually associated with functional outcome and mortality in acute stroke patients treated with tissue-type plasminogen activator. Considering profiles of NIHSS subscores rather than only the total score is informative for prognostication, particularly for assessments collected 24 hours after stroke onset.

A novel item anomaly detection approach against shilling attacks in collaborative recommendation systems using the dynamic time interval segmentation technique

Various types of web applications have gained both higher customer satisfaction and more benefits since being successfully armed with personalized recommendation. However, the increasingly rampant shilling attackers apply biased rating profiles to systems to manipulate item recommendations, which not just lower the recommending precision and user satisfaction but also damage the trustworthiness of intermediated transaction platforms and participants. Many studies have offered methods against shilling attacks, especially user profile based-detection. However, this detection suffers from the extraction of the universal feature of attackers, which directly results in poor performance when facing the improved shilling attack types. This paper presents a novel dynamic time interval segmentation technique based item anomaly detection approach to address these problems. In particular, this study is inspired by the common attack features from the standpoint of the item profile, and can detect attacks regardless of the specific attack types. The proposed segmentation technique could confirm the size of the time interval dynamically to group as many consecutive attack ratings together as possible. In addition, apart from effectiveness metrics, little attention has been paid to the robustness of detection methods, which includes measuring both the accuracy and the stability of results. Hence, we introduced a stability metric as a complement for estimating the robustness. Thorough experiments on the MovieLens dataset illustrate the performance of the proposed approach, and justify the value of the proposed approach for online applications.

National institutes of health stroke scale item profiles as predictor of patient outcome: external validation on independent trial data.

National Institutes of Health Stroke Scale (NIHSS) item profiles that were recently proposed may prove useful both clinically and for research studies. We aimed to validate the NIHSS item profiles in an acute cohort. We conducted a retrospective analysis on pooled data from randomized clinical trials. We applied the latent class analysis probabilities of profile membership developed from the derivation study to obtain symptom grouping, a-NIHSS item profiles. We implemented an independent latent class analysis to derive secondary symptom grouping, b-NIHSS item profiles. Validation was performed by assessing the associations with outcomes and evaluating both sets of NIHSS item profiles' discrimination and calibration to the data. The outcomes evaluated included modified Rankin Scale (mRS; using the full distribution and dichotomized, mRS, 0-1) at day 90 and mortality by 90 days. We identified 10 271 patients. Ordinal analysis of mRS confirmed increased odds of better outcome across the profiles in a stepwise manner, adjusted for age and thrombolysis treatment, for each set of NIHSS item profiles. Similar patterns were observed for mRS 0 to 1, and inverse patterns were seen for mortality. The c-statistics of a-NIHSS and b-NIHSS item profiles for mRS 0 to 1 were similar at 0.71 (95% confidence interval, 0.70-0.72) and for mortality, 0.74 (0.73-0.75) and 0.75 (0.73-0.76), respectively. Calibration was good. These NIHSS item profiles identified using latent class analysis offer a reliable approach to capture the true response patterns that are associated with functional and outcome and mortality post stroke. This approach has the potential to enhance the clinical value of the overall NIHSS score.

What is the relationship between the lexical profile of test items and performance on a standardized English proficiency test?

Lexical profiling research has indicated the vocabulary sizes that may be necessary to comprehend different spoken and written text types. These figures are based on studies of lexical coverage that have indicated that comprehension is likely to rise as the number of known words in a text increases (e.g., Laufer, 1989). This study examined the lexical profiles of passages included in an English L2 proficiency test used for university admission purposes. A total of 87 reading comprehension, listening comprehension, and cloze passages from CanTEST were analyzed to determine the vocabulary size needed to reach 95% and 98% lexical coverage. The results indicated that there was large variation between the lexical profiles of the texts. At the 1000 word frequency level, there were differences ranging from 15.22% to 20.05% coverage between the most and least lexical demanding passages in the three parts of the test. The correlations between the lexical profiles of the texts at the different word frequency levels and performance on the corresponding test items were calculated to determine the relationship between these two variables. The results indicated that there was either no correlation or a small correlation in all comparisons.

A Note on Parameter Estimate Comparability: Across Latent Classes in Mixture IRT Modeling.

The use of mixture item response theory modeling is exemplified typically by comparing item profiles across different latent groups. The comparisons of item profiles presuppose that all model parameter estimates across latent classes are on a common scale. This note discusses the conditions and the model constraint issues to establish a common scale across latent classes.

A tag-based recommender system using rule-based collaborative profile enrichment

With the rapid increasing rate of the high volume of social web contents due to the growing popularity of social media services, significant attention has been drawn towards recommender systems i.e. systems, that offer recommendations to users on items appropriate to their requirements. To offer suitable recommendations, the systems need comprehensive user and item models that would be able to provide thorough understanding of their characteristics and preferences. In this article, a new recommender system is proposed based on the similarities between user and item profiles. The approach here is to generate user and item profiles by discovering frequent user-generated tag patterns, and to enrich each individual profile by a two-phase profile enrichment procedure. The profiles are extended by association rules discovered through the association rule mining process. The user/item profiles are further enriched through collaboration with other similar user/item profiles. To evaluate the performance of this proposed approach, a real dataset from The Del.icio.us website is used for empirical experiment. Experimental result s demonstrate that the proposed approach provides a better representation of user interests and achieves better recommendation results in terms of precision and ranking accuracy as compared to existing methods.

Folksonomy-based user interest and disinterest profiling for improved recommendations: An ontological approach

Social tagging has revolutionized the social and personal experience of users across numerous web platforms by enabling the organizing, managing, sharing and searching of web data. The extensive amount of information generated by tagging systems can be utilized for recommendation purposes. However, the unregulated creation of social tags by users can produce a great deal of noise and the tags can be unreliable; thus, exploiting them for recommendation is a nontrivial task. In this study, a new recommender system is proposed based on the similarities between user and item profiles. The approach applied is to generate user and item profiles by discovering tag patterns that are frequently generated by users. These tag patterns are categorized into irrelevant patterns and relevant patterns which represent diverse user preferences in terms of likes and dislikes. Furthermore, presented here is a method for translating these tag-based profiles into semantic profiles by determining the underlying meaning(s) of the tags, and mapping them to semantic entities belonging to external knowledge bases. To alleviate the cold start and overspecialization problems, semantic profiles are enriched in two phases: (a) using a semantic spreading mechanism and then (b) inheriting the preferences of similar users. Experiment indicates that this approach not only provides a better representation of user interests, but also achieves a better recommendation result when compared with existing methods. The performance of the proposed recommendation method is investigated in the face of the cold start problem, the results of which confirm that it can indeed remedy the problem for early adopters, hence improving overall recommendation quality.

Calorie Reductions and Within-Meal Calorie Compensation in Children's Meal Combos

Objective: What happens if a fast-food restaurant chain lowers the calories in a children’s meal bundle (CMB), mainly by reducing the portion size of French fries? This study examines how such changes may influence within-meals selections. Specifically, do lower-calorie changes lead to within-meal calorie compensation? Design and Methods: Item-level anonymous transaction data were collected for thirty representative US restaurants during the months of June, July and August of 2011 (pre-changes) and 2012 (post-changes) with a focus on transaction records that included a CMB. Mixed effects, repeated measures estimation techniques were used for analysis. Outcome measures were the number and caloric profile of specific entree items, side items, and beverages purchased in all children’s meals. Results: Based on transaction data, the new CMB resulted in children’s meals that had an average of 18.8% fewer calories (p Conclusion: Small changes in the automatic – or default – foods offered or promoted in children’s meals can reduce calorie intake and improve the overall nutrition from selected foods as long as there is still an indulgence. Importantly, under these conditions, small enough changes do not appear to encourage within-meal calorie compensation among children.

A cross‐language personalized recommendation model in digital libraries

PurposeThe purpose of this paper is to develop a cross‐language personalized recommendation model based on web log mining, which can recommend academic articles, in different languages, to users according to their demands.Design/methodology/approachThe proposed model takes advantage of web log data archived in digital libraries and learns user profiles by means of integration analysis of a user's multiple online behaviors. Moreover, keyword translation was carried out to eliminate language dissimilarity between user and item profiles. Finally, article recommendation can be achieved using various existing algorithms.FindingsThe proposed model can recommend articles in different languages to users according to their demands, and the integration analysis of multiple online behaviors can help to better understand a user's interests.Practical implicationsThis study has a significant implication for digital libraries in non‐English countries, since English is the most popular language in current academic articles and it is a very common phenomenon for users in these countries to obtain literatures presented by more than one language. Furthermore, this approach is also useful for other text‐based item recommendation systems.Originality/valueA lot of research work has been done in the personalized recommendation area, but few works have discussed the recommendation problem under multiple linguistic circumstances. This paper deals with cross‐language recommendation and, moreover, the proposed model puts forward an integration analysis method based on multiple online behaviors to understand users' interests, which can provide references for other recommendation systems in the digital age.

Improved Collaborative Filtering using Evolutionary Algorithm based Feature Extraction

ubiquity of Collaborative Filtering systems is evident in the wide variety of domains to which they have been applied successfully. However a major challenge to such systems is the high dimensionality and sparsity of the expressed preferences. Dealing effectively with large user profiles would improve the scalability of the system whereas reducing sparsity would increase the quality of recommendations. Several approaches in this direction have focused on feature selection and feature extraction in order to reduce the data dimension and thus make the recommendation process more scalable. Some of the features extraction techniques are based on extracting content based features. However many such solutions have been handcrafted and thus not guaranteed to work optimally under all data environments. This work explores Evolutionary algorithms based feature extraction techniques where the extracted features may describe user or item profiles. The features constructed/extracted thus are compact, dense and are discriminative. Moreover they have the advantage of requiring no extra information (such as content description) and are adaptive, delivering the optimal feature extraction scheme for the particular dataset. We have performed experiments with the popular MovieLens dataset and have compared the user-based and item-based evolutionary feature extraction schemes with respect to their accuracy. The experiments establish that the evolutionary feature extraction schemes score over traditional algorithms as well as content-based feature extraction schemes. Moreover we find that the item-based evolutionary feature extraction schemes outperform their user-based counterparts under varying parameter values.

Effects of phylogeny and prey type on fatty acid calibration coefficients in three pinniped species: implications for the QFASA dietary quantification technique

Quantitative fatty acid signature analysis (QFASA) has been proposed as a technique for determining the long-term diet of animals. The method compares the fatty acid (FA) profiles of predators and potential prey items to estimate relative prey intake. We tested the assumptions of a key step in QFASA, the correction of predator FA signatures for metabolic processes through sets of calibration coefficients (CCs). We conducted long-term controlled feeding studies with captive Steller sea lions consuming herring and eulachon and northern fur seals consuming herring. We compared the results with data from harbour seals eating herring to evaluate the effects of phylogeny and prey type on individual CCs. Even within the limited extended dietary FA subset recommended for use by other researchers, we found that at least 41% of the CCs differed by family (otariid vs. phocid seals) and 58% differed by predator species (sea lion vs. fur seal), suggesting that CCs may be highly species-specific. We also found that 64% of the CCs differed by prey type (sea lions consuming herring vs. eulachon), which raises some fundamental implementation issues. We also found significant differences in diet predictions when the herringand eulachon-derived sets of CCs were applied to an actual multi-species diet. CCs are presently used as a simple mathematical attempt to describe potentially complex biochemistry. The results of this study raise questions regarding the validity of using CCs derived from an alternative predator species, and highlight some fundamental issues regarding QFASA methodology that need to be addressed through further controlled studies.

Binomial Real Option Pricing for Restaurant Menu Analysis

Numerous researchers have proposed methods to improve the profitability of restaurant menus. The profit-analysis techniques that have been proposed generally assess menu items according to popularity as against comparative contribution margin or a similar scheme that theoretically allows managers to apply a portfolio adjustment approach or simply to replace items that are low in profitability. However, none of the analytical models discussed in the literature consider the element of time, which is an essential consideration in planning for menu item replacement. This study uses a real option pricing model to construct time-axis profiles of menu items and to analyze their potential profitability given varying time constraints. The analysis applies a binomial real option pricing model, as applied to actual data, and a real option pricing model was used to compare the advantages and benefits of different menu items. A comparison of the binomial model’s results with those of other menu analysis models found that the time-based model was more effective in determining how to construct a menu portfolio.

Nutrient-based standards for school lunches complement food-based standards and improve pupils' nutrient intake profile

Following concerns about the nutritional content of school lunches and the increased prevalence of overweight and obesity in the UK, changes to the standards of school meals were made. From September 2008, all primary schools in England were required, by law, to be fully compliant with the new food-based standards (FBS) and nutrient-based standards (NBS) for school lunches. The aim of the present survey was to evaluate the introduction of the NBS for school lunches on the nutritional profile of food and drink items provided by schools and chosen by pupils at lunchtime. A nationally representative sample of 6696 pupils from 136 primary schools in England aged 3-12 years and having school lunches was recruited. Data were collected on lunchtime food and drink provision at each school and on pupil food and drink choices at lunchtime. Caterers also provided planned menus, recipes and other cooking information. Compliance with both the FBS and NBS was then assessed. Results show that even when the FBS was met, many schools did not provide a school lunch that met the NBS as well. The average school lunch eaten was significantly lower in fat, saturated fat and Na in schools that met both the FBS and NBS for school lunches compared with schools that met only the FBS. Change in school lunch policy has contributed to improvements in pupils' choices and the nutritional profile of foods selected at lunchtime.

Ontological content‐based filtering for personalised newspapers

PurposeThe purpose of this paper is to describe a new ontological content‐based filtering method for ranking the relevance of items for readers of news items, and its evaluation. The method has been implemented in ePaper, a personalised electronic newspaper prototype system. The method utilises a hierarchical ontology of news; it considers common and related concepts appearing in a user's profile on the one hand, and in a news item's profile on the other hand, and measures the “hierarchical distances” between these concepts. On that basis it computes the similarity between item and user profiles and rank‐orders the news items according to their relevance to each user.Design/methodology/approachThe paper evaluates the performance of the filtering method in an experimental setting. Each participant read news items obtained from an electronic newspaper and rated their relevance. Independently, the filtering method is applied to the same items and generated, for each participant, a list of news items ranked according to relevance.FindingsThe results of the evaluations revealed that the filtering algorithm, which takes into consideration hierarchically related concepts, yielded significantly better results than a filtering method that takes only common concepts into consideration. The paper determined a best set of values (weights) of the hierarchical similarity parameters. It also found out that the quality of filtering improves as the number of items used for implicit updates of the profile increases, and that even with implicitly updated profiles, it is better to start with user‐defined profiles.Originality/valueThe proposed content‐based filtering method can be used for filtering not only news items but items from any domain, and not only with a three‐level hierarchical ontology but any‐level ontology, in any language.

Factors Associated with Parent–Child (Dis)agreement on Child Behavior and Parenting Problems in Chinese Immigrant Families

We examined familial and cultural factors predicting parent–child (dis)agreement on child behavior and parenting problems. Immigrant Chinese parents (89.7% mothers; M age = 44.24 years) and their children (62 boys; 57.9%) between the ages of 9 and 17 years (M = 11.9 years, SD = 2.9) completed measures of parent punitive behavior and child problems. Concordance in item profiles and discrepancies in overall problem levels were assessed. Overall, immigrant parents reported fewer child and parenting problems than did their children. Relationship closeness predicted less disagreement in ratings of child internalizing symptoms and punitive parenting. Parental acculturative stress and parent–child acculturation dissonance predicted more disagreement regarding internalizing problems. The findings highlight potential under-identification of internalizing problems among immigrant Chinese families that may be driven by acculturation processes.

Establishing the phraseological profile of a text type

Starting from the observation that meaning does not primarily reside in individual words but in the phrase, this paper focuses on the examination of recurring phrases in language. It introduces a new analytical model that leads corpus researchers to a profile of the central phraseological items in a selected text or text collection. In this paper, the model is applied to a 3.5-million word corpus of online academic book reviews that represents part of the specialized discourse of the global community of linguists. This demonstrates how the model facilitates the study of the occurrence and distribution of the central phraseological items in linguistic book reviews, and how it helps to determine the extent of the phraseological tendency of language.