Ranking Measures Research Articles

Context based Adoption of Ranking and Indexing Measures for Cricket Team Ranks

Context based Adoption of Ranking and Indexing Measures for Cricket Team Ranks

Graph-based rank aggregation method for high-dimensional and partial rankings

Rank aggregation has recently become a common approach for combining individual rankings into a consensus and for quantifying and improving performance in various applications, such as elections, web page rankings, and sports. During the past few years, rankings from many sources have become increasingly high-dimensional and partial. In this study, we develop a rank aggregation method by constructing a directed weighted competition graph. We introduce the concept of “ratio of out- and in-degrees (ROID)” to transform high-dimensional partial rankings into a single consensus. Moreover, we provide a novel effectiveness measure for the aggregate ranking according to its deviations from the ground truth ranking. The proposed method is compared with four typical methods with synthetic rankings. The results indicate that our method outperforms the other four by a significant margin and can be particularly efficient in aggregating high-dimensional rankings. The empirical results validate the effectiveness and feasibility of our method.

Ranking and Similarity Measures of Interval-Valued Pentagonal Fuzzy Numbers

We define generalized interval-valued pentagonal fuzzy numbers. Based on the height of the lower and upper pentagonal fuzzy numbers we propose to categorize interval-valued pentagonal fuzzy numbers into three different categories. Using signed distance concept de-fuzzification of interval-valued pentagonal fuzzy numbers is proposed and mathematical formulas are derived using   cut representations. We also introduce two similarity measures for interval-valued pentagonal fuzzy numbers based on geometric distance, area and height of pentagonal fuzzy numbers and geometric distance, perimeter and height of pentagonal fuzzy numbers respectively. Algorithms for finding de-fuzzification value and similarity measures are proposed with flow-chart illustration.

Sentimental Analysis And Lsi Similarity Measure For Efficient Page Ranking

Supplementary factor to the general pagerank calculation which is utilized by Google chrome to rank sites in their web index results is tended to in this paper. These extra factors incorporate couple of ideas which expressly results to build the precision of evaluating the PageRank value. By making a decision about the likeness between the web page content with the text extracted from different site pagesresulted in topmost search using few keywords of the considered page for which the rank is to be determined by utilizing a comparability measure. It results with a worth or rate which speaks to the significance or similarity factor. Further, in a similar strategy if sentimental analysis is applied the search results of the keywords could be analysed with keywords of the page considered, it results with a Sentimental Analysed factor.In this way, one can improve and execute the Page ranking procedure which results with a superior accuracy.Hadoop Distributed File System is used to compute the page rank of input nodes. Python is chosen for parallel page rank algorithm that is executed on Hadoop.

What Should Law School Rankings Measure and How Should We Measure it: A Comment on Heald & Sichelman's Rankings

There are obvious benefits to ranking academic departments based on objective measures of faculty research output. However, there are considerable difficulties associated with producing reliable and accurate rankings. In this short comment, we offer an evaluation of Heald and Sichelman's recent foray into the project of ranking law schools. Heald and Sichelman are to be commended for the transparency and rigor of their rankings effort. At the same time, it is important to note that their rankings involve a series of contestable discretionary choices and could give rise to potential counter-productive gaming by law schools seeking to improve their place in the rankings. In particular, Heald and Sichelman's system places a thumb on the scale on behalf of more senior faculty who publish in traditional law reviews and write in popular substantive areas like constitutional law. This raises the concern that rankings of this type could discourage law schools from hiring faculty that are young, produce interdisciplinary scholarship, and write in otherwise under-represented fields. Nonetheless, Heald and Sichelman have taken an important step forward, and other scholars should look to build profitably upon their work.

Investigation of the Moderating Effect of Culture on Users’ Susceptibility to Persuasive Features in Fitness Applications

Persuasive technologies have been identified as a potential motivational tool to tackle the rising problem of physical inactivity worldwide, with research showing they are more likely to be successful if tailored to the target audience. However, in the physical activity domain, there is limited research on how culture moderates users’ susceptibility to the various persuasive features employed in mobile health applications aimed to motivate behavior change. To bridge this gap, we conducted an empirical study among 256 participants from collectivist (n = 67) and individualist (n = 189) cultures to determine their culture-specific persuasion profiles with respect to six persuasive features commonly employed in fitness applications on the market. The persuasive features include two personal features (goal-setting/self-monitoring and reward) and four social features (competition, cooperation, social learning and social comparison). We based our study on the rating of storyboards (on which each of the six persuasive features is illustrated) and the ranking of the six persuasive features in terms of perceived persuasiveness. The results of our analysis showed that users from individualist and collectivist cultures significantly differ in their persuasion profiles. Based on our rating measure, collectivist users are more likely to be susceptible to all six persuasive features (personal and social) than individualist users, who are only likely to be susceptible to personal features. However, based on our ranking measure, individualist users are more likely to be susceptible to personal features (goal-setting/self-monitoring and reward) than collectivist users. In contrast, collectivist users are more likely to be susceptible to social features (cooperation and social learning) than individualist users. Based on these findings, we provide culture-specific persuasive technology design guidelines. Our study is the first to uncover the moderating effect of culture on users’ susceptibility to commonly employed persuasive features in fitness applications.

A Longitudinal Analysis of University Rankings

Pressured by globalization and the increasing demand for public organisations to be accountable, efficient and transparent, university rankings have become an important tool for assessing the quality of higher education institutions. It is therefore important to carefully assess exactly what these rankings measure. In this paper, the three major global university rankings, The Academic Ranking of World Universities, The Times Higher Education and the Quacquarelli Symonds World University Rankings, are studied. After a description of the ranking methodologies, it is shown that university rankings are stable over time but that there is variation between the three rankings. Furthermore, using Principal Component Analysis and Exploratory Factor Analysis, we show that the variables used to construct the rankings primarily measure two underlying factors: a universities reputation and its research performance. By correlating these factors and plotting regional aggregates of universities on the two factors, differences between the rankings are made visible. Last, we elaborate how the results from these analysis can be viewed in light of often voiced critiques of the ranking process. This indicates that the variables used by the rankings might not capture the concepts they claim to measure. Doing so the study provides evidence of the ambiguous nature of university ranking's quantification of university performance.

Deep Position-Sensitive Tracking

Classification-based tracking strategies often face more challenges from intra-class discrimination than from inter-class separability. Even for deep convolutional neural networks that have been widely proven to be effective in various vision tasks, their intra-class discriminative capability is still limited by the weakness of softmax loss, especially for targets not seen in the training dataset. By taking intrinsic attributes of training samples into account, in this paper, we propose a position-sensitive loss coupled with softmax loss to achieve intra-class compactness and inter-class explicitness. Particularly, two additive margins are introduced to encode the position attribute for decision boundary maximization, which is also utilized with the proposed loss to supervise the fine-tuned features on the pre-trained model. With the nearest neighbor ranking measurement in the feature embedding domain, the whole scheme is able to reach an optimized balance between the feature-level inter-class semantic separability and instance-level intra-class relative distance ranking. We evaluate the proposed work on different popular benchmarks, and experimental results demonstrate that our tracking strategy performs favorably against most of the state-of-the-art trackers in the comparison of accuracy and robustness.

Could nature help children rise out of poverty? Green space and future earnings from a cohort in ten U.S. cities

BackgroundGrowing up in poverty is associated with poor health, and the American Dream of upward mobility is becoming an illusion for many low-income children. But nearby green space can support academic achievement, creativity, and emotional regulation, and these traits might help children rise out of poverty. ObjectivesTo examine the relationship between recent incomes of children born into poverty in the ten largest U.S. cities and densities of residential green space during childhood. MethodsWe calculate park proximity, park acreage, new park development, and NDVI greenness for 1980–1990 from Landsat and Trust for Public Land data. We obtain the 2014 income for children born between 1978 and 1982 into families in poverty from The Opportunity Atlas cohort, aggregated at the tract level (n = 5849). ResultsConditional autoregressive (CAR) models of tracts show statistically significant associations between income rank and above-average levels of greenness but not between income rank and park measures, adjusting for individual and neighborhood confounders and spatial autocorrelation. We estimate that, over a 30-year career, children growing up in tracts with the most vegetative cover will earn cumulatively $28,000 more than children growing up in tracts with the least cover, on average. Tracts with lower than average levels of precipitation, higher disadvantage, higher population density, or higher annual temperatures do not show beneficial effects of green space. ConclusionsGreenness may be weakly associated with children rising out of poverty in wetter, cooler, less-dense, more advantaged census tracts.

Detailed analysis of paternal knockout Grb10 mice suggests effects on stability of social behavior, rather than social dominance.

Imprinted genes are highly expressed in monoaminergic regions of the midbrain and their functions in this area are thought to have an impact on mammalian social behaviors. One such imprinted gene is Grb10, of which the paternal allele is generally recognized as mediating social dominance behavior. However, there has been no detailed study of social dominance in Grb10 +/p mice. Moreover, the original study examined tube‐test behavior in isolated mice 10 months of age. Isolation testing favors more territorial and aggressive behaviors, and does not address social dominance strategies employed in group housing contexts. Furthermore, isolation stress impacts midbrain function and dominance related behavior, often through alterations in monoaminergic signaling. Thus, we undertook a systematic study of Grb10 +/p social rank and dominance behavior within the cage group, using a number of convergent behavioral tests. We examined both male and female mice to account for sex differences and tested cohorts aged 2, 6 and 10 months to examine any developments related to age. We found group‐housed Grb10 +/p mice do not show evidence of enhanced social dominance, but cages containing Grb10 +/p and wild‐type mice lacked the normal correlation between three different measures of social rank. Moreover, a separate study indicated isolation stress induced inconsistent changes in tube test behavior. Taken together, these data suggest future research on Grb10 +/p mice should focus on the stability of social behaviors, rather than dominance per se.

Fractal dimensions of spectral measures of rank one perturbations of a positive self-adjoint operator

Let H be a Hilbert space. Suppose A is a positive self-adjoint operator on H and φ∈H is a cyclic unit vector. For each λ∈R, we can define the rank one perturbation of A by Aλ=A+λ〈φ,⋅〉φ. To each Aλ we can consider the spectral measure of φ, which we denote by μλ. This generates a family of measures, {μλ}, and we analyze the packing dimension of this family. Past results have determined that the Hausdorff dimension of this family can be determined if the limit inferior of a ratio involving μ is constant on a Lebesgue typical set. This ratio is sometimes called the pointwise dimension of μ and is related to the upper derivative of μ. Work has been done to make a similar argument for the packing dimension, but with little success. Using the theory of rank one perturbations and Borel transforms, we introduce the concept of Lebesgue exact dimension for μ, which allows us to determine the packing dimension of spectral measures of almost every rank one perturbation μλ. If the Lebesgue exact dimension for μ is 1<α<2 then the packing dimension of Lebesgue almost every μλ is 2−α. As a corollary, we find that this limit condition implies a stronger result: the Hausdorff and packing dimensions are equal for almost every μλ.

Modeling Side Information in Preference Relation based Restricted Boltzmann Machine for recommender systems

A majority of the collaborative filtering techniques exploit user-item rating information to generate recommendations of unseen items for a user. However, a user’s preference also depends on some extra information like item features, user attributes and others, which is known as side information. Further, according to recent studies, using preference relation as an alternative to absolute ratings often produces quality recommendations. This study proposes a collaborative filtering technique using Preference Relation based Restricted Boltzmann Machine for recommender system. The proposed method takes the preference relations of items as input and generates a ranking of items for any user. Using Conditional Restricted Boltzmann Machine, the side information of items along with preference relations are integrated into the model. Besides side information, the proposed method is also able to capture second order and higher order user-item interactions. Experimental verification of the proposed model is done using three datasets: MovieLens-1M, MovieLens-20M, and Book-Crossing, which are the most widely used datasets for testing recommender systems. Results obtained at different positions using standard ranking measures like, NDCG and MAP, indicate that the performance of the proposed method is better compared to related state-of-the-art collaborative filtering models for Top-N recommendation task.

Adaptive nonnegative matrix factorization and measure comparisons for recommender systems

The Nonnegative Matrix Factorization (NMF) of the rating matrix has shown to be an effective method to tackle the recommendation problem. In this paper we propose new methods based on the NMF of the rating matrix and we compare them with some classical algorithms such as the SVD and the regularized and unregularized non-negative matrix factorization approach. In particular a new algorithm is obtained changing adaptively the function to be minimized at each step, realizing a sort of dynamic prior strategy. Another algorithm is obtained modifying the function to be minimized in the NMF formulation by enforcing the reconstruction of the unknown ratings toward a prior term. We then combine different methods obtaining two mixed strategies which turn out to be very effective in the reconstruction of missing observations. We perform a thoughtful comparison of different methods on the basis of several evaluation measures. We consider in particular rating, classification and ranking measures showing that the algorithm obtaining the best score for a given measure is in general the best also when different measures are considered, lowering the interest in designing specific evaluation measures. The algorithms have been tested on different datasets, in particular the 1M, and 10M MovieLens datasets containing ratings on movies, the Jester dataset with ranting on jokes and Amazon Fine Foods dataset with ratings on foods. The comparison of the different algorithms, shows the good performance of methods employing both an explicit and an implicit regularization scheme. Moreover we can get a boost by mixed strategies combining a fast method with a more accurate one.

Comprehensive evaluation of h-index and its extensions in the domain of mathematics

Until the late 90’s, conventional bibliometric indices such as, publication count, citation count, and number of co-authors have extensively been employed by the scientific community to rank the researchers. In 2005, inception of a renowned ranking measure h-index has grabbed the substantial importance; the community considered h-index as a quintessential ranking measure. Subsequently, different variants and extensions of h-index have also been proposed. To date, plethora of the studies exists that harnesses h-index, its variants and extensions for researchers’ ranking. Nonetheless, the community does not agree upon a single measure that can be deemed as an optimal ranking measure. This is due to the fact that most of the contemporary studies have evaluated them either by employing a small amount of data or presumed their significance on the basis of hypothetical or imaginary case scenarios. We argue that comprehensive empirical investigation of these measures must be performed in order to tackle their real behavior. This study evaluates the h-index and its extensions by employing a comprehensive data set of authors from Mathematics discipline. The first experimental step involves the computation of correlation among the obtained values of the extensions to determine the similarity and divergence among them. Afterwards, we considered the data of international award winners from four prestigious Mathematics societies as benchmark to validate the potential of these measures by analyzing the dependence of societies on them. The outcomes revealed that overall 45% of the authors have appeared at the top occurrences of the ranked list. Out of all extensions, fraction count on paper has outperformed by bringing 55% of the awardees at top 10% of its ranked list.

A Novel Approach to Rank Text-based Essays using Pagerank Method Towards Student’s Motivational Element

Learning outcomes is one of the important factors to measure student achievement during the learning process. Today’s learning is more focused on problem-solving and reasoning to existing problems than an ordinary problem. Most exams have been directed to analysis questions for students to think and synthesize. As such is troublesome for most students, they are not ready to answer the question, thus, their answers almost similar to their friends. This implies that the teacher has tried to guide students to work professionally and originally. However, the Teacher facing difficulty assessing student’s work, especially if the assignments are conducted online without face-to-face instructions/discussions. To bridge such a gap, the teacher needs a method or algorithm to measure their rank to encourage students making an original answer. This research provides a solution in calculating students' ranks based on the similarity score of the essay answers. Pagerank is a ranking algorithm used by Google, this algorithm utilizes a Markov matrix that contains the direction of similarity score for each student. These scores are computed by the power method until converging. Rank is displayed to the teacher to review the similarity level of students' answers. As such is presented a line chart in which the x-axis refers to the students and the y-axis depicts the level of similarity. Ranking computation in Matlab produces an Eigen vector which acts as the rank measure. The higher the rank, the more similar is their answers to others. Hence, students with high ranks to work their answers more seriously ensure their original thoughts. In conclusion, the similarity score matrix using the PageRank algorithm can contribute to the teacher in providing peer motivation and encouraging student’s internal motivation by presenting the ranked-answers presentation.

What Do Rankings Measure? The U.S. News Rankings and Student Experience at Liberal Arts Colleges

What Do Rankings Measure? The U.S. News Rankings and Student Experience at Liberal Arts Colleges

Correlation analysis: edge betweenness centrality vs. neighbourhood overlap

We explore the correlation between two well-known edge centrality metrics: a computationally-heavy metric called edge betweenness centrality (EBWC) and a computationally-light metric called neighbourhood overlap (NOVER). Three different correlation measures are used for the analysis: Spearman's correlation measure (for network-wide ranking), Kendall's correlation measure (for pair-wise relative ordering) and Pearson's correlation measure (for regression-based prediction). A suite of 47 real-world networks of diverse degree distributions have been used for the correlation study. Results of the correlation analysis indicate that NOVER could be used as a computationally-light alternative to rank the edges (network-wide) in lieu of EBWC, but might not be appropriate to predict the actual EBWC values.

Correlation analysis: edge betweenness centrality vs. neighbourhood overlap

We explore the correlation between two well-known edge centrality metrics: a computationally-heavy metric called edge betweenness centrality (EBWC) and a computationally-light metric called neighbourhood overlap (NOVER). Three different correlation measures are used for the analysis: Spearman's correlation measure (for network-wide ranking), Kendall's correlation measure (for pair-wise relative ordering) and Pearson's correlation measure (for regression-based prediction). A suite of 47 real-world networks of diverse degree distributions have been used for the correlation study. Results of the correlation analysis indicate that NOVER could be used as a computationally-light alternative to rank the edges (network-wide) in lieu of EBWC, but might not be appropriate to predict the actual EBWC values.

Multi-Label Question Classification for Factoid and List Type Questions in Biomedical Question Answering

Biomedical experts and bio-curators are unable to quickly find short and precise information using typical search engines as the amount of biomedical literature is increasing exponentially. The research community is focusing on biomedical question answering (QA) systems so that anyone can find precise information nuggets from the massive amount of biomedical literature. Generally, the user queries fall under different categories such as factoid, list, yes/no, or summary. The existing state-of-the-art question answering systems deal with most of these question types. However, the research to improve the performance of individual question types is also on the rise. To improve QA system performance, question classification plays a vital role for factoid and list type questions as it allows the answer processing stage to narrow down the candidate answer space and assigns a higher rank to the correct answers. A single biomedical answer or entity may be associated with more than one biomedical category or semantic type, e.g., Coenzyme Q(10) is classified under two categories in Unified Medical Language System (UMLS): organic chemical and biologically active substance . This inherent characteristic of biomedical entities makes question classification in the biomedical domain a multi-label classification problem, where one question might expect answers belonging to more than one semantic type. To the best of our knowledge, several QA systems deal with question classification as a multi-class classification problem and only one state-of-the-art system – OAQA – deals with it as a multi-label classification problem. In this paper, we analyze the pipeline of the OAQA system for factoid and list type questions, emphasizing the multi-label question classification. We use an improved question classification dataset with the copy transformation technique to improve the performance of list type questions. Moreover, we introduce a binary transformation in the pipeline of factoid questions to increase its performance. Our modified methodology enhances the performance of both list and factoid type questions by a margin of 2% and 3% evaluated on standard $F_{1}$ and Mean Reciprocal Rank measures, respectively.

Social rank theory of depression: A systematic review of self-perceptions of social rank and their relationship with depressive symptoms and suicide risk.

Depression can be debilitating, as well as a risk factor for self-harm and suicide. Social rank theory (SRT) suggests depression stems from feelings of defeat and entrapment that ensue from experiencing oneself to be of lower rank than others. This study aims to review the literature investigating the relationship between self-perceptions of social rank and depressive symptoms or suicidal ideation/behaviour. A keyword search of three psychological and medical databases was completed (Psychinfo, Medline, Web of Knowledge). Studies were quality assessed using established criteria. An initial 1290 records were identified. After application of inclusion and exclusion criteria, 70 remained measuring depressive symptoms (n = 68), self-harm (n = 3) and suicidal ideation (n = 3). The main measures assessing social rank were the social comparison scale (SCS; n = 32) and subjective social status (SSS, n = 32), with six additional papers including another measure of social rank. In univariate analyses, as perceptions of social rank decreased, depressive symptoms (and suicidal ideation/self-harm) increased. Multivariate analyses indicated that social rank may act as a psychosocial mechanism to explain the relationship between social factors (in particular socio-economic status) and depressive symptoms. Additionally, psychological variables, such as rumination or self-esteem, may mediate or moderate the relationship between social rank and depressive or suicidal symptoms. Study quality was variable and 89% of studies were cross-sectional. Although more prospective research is required, this review highlights the importance of understanding an individual's perception of their social position compared to others as it may lead to an enhanced understanding of the aetiology of depressive disorders.