Search results

Purpose — Ranking is a natural task for a search engine; a search engine result page is the most common example. This chapter aims at illustrating the motivations and the concepts of rank correlation in a practical way for the researchers active in the different domains of search engines.

Methodology/approach — To this end, this chapter provides a survey according to a topic-oriented basis of the search engine evaluation literature specifically devoted to or based on rank correlation; the chapter explains and illustrates how statistics is the only approach to rank correlation.

Findings/research limitations/implications — The chapter introduces the pros and cons of rank correlation measures through a light-weight formal description and a number of concrete examples to find the measure that better fit a context.

Practical implications — This chapter provides a blueprint for the application of rank correlation within scientific experimentation or item/service recommendation.

Social implications — Rank correlation analyses impact on the success or failure of a search engine in performing the tasks for which it has been designed and hence on the people's daily life activities.

Originality/value of paper — This chapter places rank correlation within a scientific research perspective and in particular connects to and complements documentation on search engine evaluation.

We study the statistical properties of Pearson correlation coefficients of Gaussian ranks, and Gaussian rank regressions – ordinary least-squares (OLS) models applied to those ranks. We show that these procedures are fully efficient when the true copula is Gaussian and the margins are non-parametrically estimated, and remain consistent for their population analogs otherwise. We compare them to Spearman and Pearson correlations and their regression counterparts theoretically and in extensive Monte Carlo simulations. Empirical applications to migration and growth across US states, the augmented Solow growth model and momentum and reversal effects in individual stock returns confirm that Gaussian rank procedures are insensitive to outliers.

The application of conventional rank correlation tests to the comparison of relevance judgements is outlined and their disadvantages are noted. A weighted rank correlation coefficient rw is proposed, which assigns variable weights to exchanges of rank on different parts of the scale. Examples of its computation are given, both for tied and untied ranks, and the values of rw for n=3 and n=4 are tabulated. The chief mathematical properties of the coefficient are investigated.

What copulas are, their estimation, and use is illustrated using a geographical diversification example. To accomplish this, dependencies between county-level yields are calculated for non-irrigated wheat, upland cotton, and sorghum using Pearson linear correlation and Kendall's tau. The use of Kendall's tau allows the implementation of copulas to estimate the dependency between county-level yields. The paper aims to discuss these issues.

Four parametric copulas, Gaussian, Frank, Clayton, and Gumbel, are used to estimate Kendall's tau. These four estimates of Kendall's tau are compared to Pearson's linear correlation, a more typical measure of dependence. Using this information, functions are estimated to determine the relationships between dependencies and changes in geographic and climate data.

The effect on county-level crop yields based on changes of geographical and climate variables differed among the different dependency measures among the three different crops. Implementing alternative dependency measures changed the statistical significance and the signs of the coefficients in the sorghum and cotton dependence functions. Copula-based elasticities are consistently less than the linear correlation elasticities for wheat and cotton. For sorghum, however, the copula-based elasticities are generally larger. The results indicate that one should not take the issue of measuring dependence as a trivial matter.

This research not only extends the current literature on geographical diversification by taking a more detailed examination of factors impacting yield dependence, but also extends the copula literature by comparing estimation results using linear correlation and copula-based rank correlation.

We investigate the question of whether investing in a child’s development by having a parent stay at home when the child is young is correlated with the child’s adult outcomes. Specifically, do children with stay-at-home mothers have higher adult earnings than children raised in households with a working mother? The major contribution of our study is that, unlike previous studies, we have access to rich longitudinal data that allows us to measure both the parental earnings when the child is very young and the adult earnings of the child. Our findings are consistent with previous studies that show insignificant differences between children raised by stay-at-home mothers during their early years and children with mothers working in the market. We find no impact of maternal employment during the first five years of a child’s life on earnings, employment, or mobility measures of either sons or daughters. We do find, however, that maternal employment during children’s high school years is correlated with a higher probability of employment as adults for daughters and a higher correlation between parent and daughter earnings ranks.

Although there has been a great deal of interest in the idea of accounting for human resources and considerable theoretical discussion of the problems of measuring human resource value and cost, there has been virtually no empirical research on the validity of proposed methods and models. This paper reports some preliminary evidence on the validity of selected surrogates of a person's value to an organization. It describes a field study conducted to determine the convergent and discriminant validity of three possible surrogates of a person's value to an organization: replacement cost, compensation, and a performance measure.

The purpose of this study is to present an alternative to university ranking lists published in U.S. News & World Report, Times Higher Education, Academic Ranking of World Universities and Money Magazine. A strategy is proposed to mine a collection of university data obtained from Twitter and publicly available online academic sources to compute social media metrics that approximate typical academic rankings of US universities.

The Twitter application programming interface (API) is used to rank 264 universities using two easily collected measurements. The University Twitter Engagement (UTE) score is the total number of primary and secondary followers affiliated with the university. The authors mine other public data sources related to endowment funds, athletic expenditures and student enrollment to compute a ranking based on the endowment, expenditures and enrollment (EEE) score.

In rank-to-rank comparisons, the authors observed a significant, positive rank correlation (τ = 0.6018) between UTE and an aggregate reputation ranking, which indicates UTE could be a viable proxy for ranking atypical institutions normally excluded from traditional lists.

The UTE and EEE metrics offer distinct advantages because they can be calculated on-demand rather than relying on an annual publication and they promote diversity in the ranking lists, as any university with a Twitter account can be ranked by UTE and any university with online information about enrollment, expenditures and endowment can be given an EEE rank. The authors also propose a unique approach for discovering official university accounts by mining and correlating the profile information of Twitter friends.

Lists of journals ranked according to number of citations received are frequently used as indicators of usefulness, but little research has been carried out to test the validity of this hypothesis. On comparing lists of titles of journals ranked by citation counting with lists of the same journals, ranked according to frequency of use (using data from a survey at the National Lending Library), it was found that the rank order correlation between the two was low. This suggests that ranked lists produced by analyses of citations do not constitute valid guides for journal selection by libraries.

The purpose of this study is to evaluate whether the classification of the equity mutual fund depends on the performance measure used.

The sample for this study includes stock mutual funds for the USA, Europe and emerging market economies covering the period 2010 to 2020. Using more than 20 performance measures the results are compared using the Sharpe ratio as the reference.

The results show that performance measures based on absolute reward–risk ratios like Sortino, Treynor, etc. have similar rankings, because in general the numerator (mean excess return) is the same. However, when the authors employ other types of performance measures, results may be significantly different, especially in the case of metrics for “incremental returns”, i.e. alphas. Focussing on markets, their results show that choosing performance measures is more relevant for emerging markets.

The sample is only limited to the USA, Europe and the emerging market, and there are other performance metrics in the literature which have not been covered in this work.

The ordering of equity mutual funds depends on the measure used, specially if investors employ factor models to measure excess returns (alphas). Hence, policy formulation on disclosure of mutual fund performance should encourage the use of several metrics from different families. Investors must be aware of the different rankings made and the most appropriate metrics based on their preferences.

This paper focusses specifically on the effect that performance metrics have on relative fund performance. Previous studies have ignored alpha metrics to rank funds, which are commonly employed by investors. The authors’ study performs an analysis for three different markets considering the two main developed ones (the American and European equity markets), as well as the emerging one, largely ignored until now.

Amid rapid technological progress, the construction industry is embracing Construction 4.0, redefining work practices through emerging technologies. However, the implications of Construction 4.0 technologies to enhancing well-being are still poorly understood. Particularly, the challenge lies in selecting technologies that critically contribute to well-being enhancement. Therefore, this study aims to evaluate the implications of Construction 4.0 technologies to enhancing well-being.

A list of Construction 4.0 technologies was identified from a national strategic plan on Construction 4.0, using Malaysia as a case study. Fourteen construction industry experts were selected to evaluate the implications of Construction 4.0 technologies on well-being using fuzzy Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). The expert judgment was measured using linguistic variables that were transformed into fuzzy values. Then, the collected data was analyzed using the following analyses: fuzzy TOPSIS, Pareto, normalization, sensitivity, ranking performance and correlation.

Six Construction 4.0 technologies are critical to enhancing well-being: cloud & real-time collaboration, big data & predictive analytics, Internet of Things, building information modeling, autonomous construction and augmented reality & virtualization. In addition, artificial intelligence and advanced building materials are recommended to be implemented simultaneously as a very strong correlation exists between them.

The novelty of this study lies in a comprehensive understanding of the implications of Construction 4.0 technologies to enhancing well-being. The findings can assist researchers, industry practitioners and policymakers in making well-informed decisions to select Construction 4.0 technologies when targeting the enhancement of the overall well-being of the local construction industry.