Do-Rank: DCG optimization for learning-to-rank in tag-based item recommendation systems

Discounted Cumulative Gain (DCG) is a well-known ranking evaluation measure for models built with multiple relevance graded data. By handling tagging data used in recommendation systems as an ordinal relevance set of {negative,null,positive}, we propose to build a DCG based recommendation model. We present an efficient and novel learning-to-rank method by optimizing DCG for a recommendation model using the tagging data interpretation scheme. Evaluating the proposed method on real-world datasets, we demonstrate that the method is scalable and outperforms the benchmarking methods by generating a quality top-N item recommendation list.

Ranking of Prospective New Generation Location for a Power Network in a Deregulated System

In developing countries high rate of growth in demand of electric energy is felt, and so the addition of new generating units becomes necessary. In deregulated power systems private generating stations are encouraged to add new generations. Finding the appropriate location of new generator to be installed can be obtained by running repeated power flows, carrying system studies like analyzing the voltage profile, voltage stability, loss analysis etc. In this paper a new methodology is proposed which will mainly consider the existing network topology into account. A concept of T-index is introduced in this paper, which considers the electrical distances between generator and load nodes.This index is used for ranking significant new generation expansion locations and also indicates the amount of permissible generations that can be installed at these new locations. This concept facilitates for the medium and long term planning of power generation expansions within the available transmission corridors. Studies carried out on a sample 7-bus system, EHV equivalent 24-bus system and IEEE 39 bus system are presented for illustration purpose.

Relative ranking: A biased rating

Reviewers' ratings have become one of the most influential parameters when making a decision to purchase or rent the products or services from the online vendors. Star Rating system is the de-facto standard for rating a product. It is regarded as one of the most visually appealing rating systems that directly interact with the consumers; helping them find products they will like to purchase as well as register their views on the product. It offers visual advantage to pick the popular or most rated product. Any system that is not as appealing as star system will have a chance of rejection by online business community. This paper argues that, the visual advantage is not enough to declare star rating system as a triumphant, the success of a ranking system should be measured by how effectively the system helps customers make decisions that they, retrospectively, consider correct. This paper argues and suggests a novel approach of Relative Ranking within the boundaries of star rating system to overcome a few inherent disadvantages the former system comes with. © Springer Science+Business Media B.V. 2010.

Dynamic security assessment of power systems using structure-preserving energy functions

An application of direct methods to dynamic security assessment of power systems using structure-preserving energy functions (SPEF) is presented. The transient energy margin (TEM) is used as an index for checking the stability of the system as well as ranking the contigencies based on their severity. The computation of the TEM requires the evaluation of the critical energy and the energy at fault clearing. Usually this is done by simulating the faulted trajectory, which is time-consuming. In this paper, a new algorithm which eliminates the faulted trajectory estimation is presented to calculate the TEM. The system equations and the SPEF are developed using the centre-of-inertia (COI) formulation and the loads are modelled as arbitrary functions of the respective bus voltages. The critical energy is evaluated using the potential energy boundary surface (PEBS) method. The method is illustrated by considering two realistic power system examples.

Structured Learning for Non-Smooth Ranking losses

Learning to rank from relevance judgment is an active research area. Itemwise score regression, pairwise preference satisfaction, and listwise structured learning are the major techniques in use. Listwise structured learning has been applied recently to optimize important non-decomposable ranking criteria like AUC (area under ROC curve) and MAP(mean average precision). We propose new, almost-lineartime algorithms to optimize for two other criteria widely used to evaluate search systems: MRR (mean reciprocal rank) and NDCG (normalized discounted cumulative gain)in the max-margin structured learning framework. We also demonstrate that, for different ranking criteria, one may need to use different feature maps. Search applications should not be optimized in favor of a single criterion, because they need to cater to a variety of queries. E.g., MRR is best for navigational queries, while NDCG is best for informational queries. A key contribution of this paper is to fold multiple ranking loss functions into a multi-criteria max-margin optimization.The result is a single, robust ranking model that is close to the best accuracy of learners trained on individual criteria. In fact, experiments over the popular LETOR and TREC data sets show that, contrary to conventional wisdom, a test criterion is often not best served by training with the same individual criterion.

Evaluation of Suitable Locations for Generation Expansion in Restructured Power Systems: A Novel Concept of T-index Thukaram

In developing countries, a high rate of growth in the demand for electric energy is felt, and so the addition of new generating units becomes inevitable. In deregulated power systems, private generating stations are encouraged to add new generations. Some of the factors considered while placing a new generating unit are: availability of esources, ease of transmitting power, distance from the load centre, etc. Finding the most appropriate locations for generation expansion can be done by running repeated power flows and carrying system studies like analyzing the voltage profile, voltage stability, loss analysis, etc. In this paper a new methodology is proposed which will mainly consider the existing network topology. A concept of T-index is introduced in this paper, which considers the electrical distances between generator and load nodes. This index is used for ranking the most significant new generation expansion locations and also indicates the amount of permissible generations that can be installed at these new locations. This concept facilitates for the medium and long term planning of power generation expansions within the available transmission corridors. Studies carried out on an EHV equivalent 10-bus system and IEEE 30 bus systems are presented for illustration purposes.

Infinite push: a new support vector ranking algorithm that directly optimizes accuracy at the absolute top of the list

Ranking problems have become increasingly important in machine learning and data mining in recent years, with applications ranging from information retrieval and recommender systems to computational biology and drug discovery. In this paper, we describe a new ranking algorithm that directly maximizes the number of relevant objects retrieved at the absolute top of the list. The algorithm is a support vector style algorithm, but due to the different objective, it no longer leads to a quadratic programming problem. Instead, the dual optimization problem involves l1, ∞ constraints; we solve this dual problem using the recent l1, ∞ projection method of Quattoni et al (2009). Our algorithm can be viewed as an l∞-norm extreme of the lp-norm based algorithm of Rudin (2009) (albeit in a support vector setting rather than a boosting setting); thus we refer to the algorithm as the ‘Infinite Push’. Experiments on real-world data sets confirm the algorithm’s focus on accuracy at the absolute top of the list.

Algorithm for mining outliers in categorical data through ranking

The rapid growth in the field of data mining has lead to the development of various methods for outlier detection. Though detection of outliers has been well explored in the context of numerical data, dealing with categorical data is still evolving. In this paper, we propose a two-phase algorithm for detecting outliers in categorical data based on a novel definition of outliers. In the first phase, this algorithm explores a clustering of the given data, followed by the ranking phase for determining the set of most likely outliers. The proposed algorithm is expected to perform better as it can identify different types of outliers, employing two independent ranking schemes based on the attribute value frequencies and the inherent clustering structure in the given data. Unlike some existing methods, the computational complexity of this algorithm is not affected by the number of outliers to be detected. The efficacy of this algorithm is demonstrated through experiments on various public domain categorical data sets.

Fisheries marketing systems and consumer preferences in Puttalam District Sri-Lanka

The aim of this study was to understand the current and historic market situation for inland fish and it’s substitutes in order to identify which of the various production opportunities presented by the seasonal tank resource might have greatest relevance for marginal communities in the Dry-zone. Regional and sub-regional market networks for fish and meat products were investigated, ranking and scoring exercises used to characterise consumer demand in rain-fed areas of North West Province and secondary data sources were used to assess historic patterns of demand and supply [PDF contains 57 pages]

The outflow ranking method for weighted directed graphs

A ranking method assigns to every weighted directed graph a (weak) ordering of the nodes. In this paper we axiomatize the ranking method that ranks the nodes according to their outflow using four independent axioms. Besides the well-known axioms of anonymity and positive responsiveness we introduce outflow monotonicity – meaning that in pairwise comparison between two nodes, a node is not doing worse in case its own outflow does not decrease and the other node’s outflow does not increase – and order preservation – meaning that adding two weighted digraphs such that the pairwise ranking between two nodes is the same in both weighted digraphs, then this is also their pairwise ranking in the ‘sum’ weighted digraph. The outflow ranking method generalizes the ranking by outdegree for directed graphs, and therefore also generalizes the ranking by Copeland score for tournaments.

Health systems efficiency after the crisis in the OECD

This work evaluates the efficiency position of the health system of each OECD country. It identifies whether, or not, health systems changed in terms of quality and performance after the financial crisis. The health systems performance was calculated by fixed-effects estimator and by stochastic frontier analysis. The results suggest that many of those countries that the crisis affected the most are more efficient than the OECD average. In addition, some of those countries even managed to reach the top decile in the efficiency ranking. Finally, we analyze the stochastic frontier efficiency scores together with other health indicators to evaluate the health systems’ overall adjustments derived from the crisis.

Divulgación on-line de los códigos de buen gobierno en las IES y su determinación para el posicionamiento en el ranking QS

Dado el interés que se presenta con los temas de gobierno corporativo, este trabajo busca describir si la divulgación on-line de los contenidos de los códigos de buen gobierno, es determinante en el posicionamiento que tienen las Instituciones de Educación Superior (IES) en el ranking QS. Partiendo de una muestra de 20 IES, se recolectaron un conjunto de datos dicotómicos para 30 variables independientes y se relacionaron con la variable dependiente denominada posicionamiento en el ranking. A partir de lo anterior, se elaboró un trabajo descriptivo y correlacional con el fin de probar las hipótesis de investigación. Este estudio reveló que la divulgación on-line de los contenidos de los códigos de buen gobierno en las IES, no es determinante para el posicionamiento en el ranking QS.

xhRank: ranking entities on the Semantic Web

In general, ranking entities (resources) on the Semantic Web (SW) is subject to importance, relevance, and query length. Few existing SW search systems cover all of these aspects. Moreover, many existing efforts simply reuse the technologies from conventional Information Retrieval (IR), which are not designed for SW data. This paper proposes a ranking mechanism, which includes all three categories of rankings and are tailored to SW data.

Unravelling the rationale of 'overgrazing' and stocking rates in the beef production systems of Central Brazil using a bi-criteria compromise programming model

In Central Brazil, the long-term sustainability of beef cattle systems is under threat over vast tracts of farming areas, as more than half of the 50 million hectares of sown pastures are suffering from degradation. Overgrazing practised to maintain high stocking rates is regarded as one of the main causes. High stocking rates are deliberate and crucial decisions taken by the farmers, which appear paradoxical, even irrational given the state of knowledge regarding the consequences of overgrazing. The phenomenon however appears inextricably linked with the objectives that farmers hold. In this research those objectives were elicited first and from their ranking two, ‘asset value of cattle (representing cattle ownership)' and ‘present value of economic returns', were chosen to develop an original bi-criteria Compromise Programming model to test various hypotheses postulated to explain the overgrazing behaviour. As part of the model a pasture productivity index is derived to estimate the pasture recovery cost. Different scenarios based on farmers' attitudes towards overgrazing, pasture costs and capital availability were analysed. The results of the model runs show that benefits from holding more cattle can outweigh the increased pasture recovery and maintenance costs. This result undermines the hypothesis that farmers practise overgrazing because they are unaware or uncaring about overgrazing costs. An appropriate approach to the problem of pasture degradation requires information on the economics, and its interplay with farmers' objectives, for a wide range of pasture recovery and maintenance methods. Seen within the context of farmers' objectives, some level of overgrazing appears rational. Advocacy of the simple ‘no overgrazing' rule is an insufficient strategy to maintain the long-term sustainability of the beef production systems in Central Brazil.

Unravelling the rationale of 'overgrazing' and stocking rates in the beef production systems of Central Brazil using a bi-criteria compromise programming model

In Central Brazil, the long-term, sustainability of beef cattle systems is under threat over vast tracts of farming areas, as more than half of the 50 million hectares of sown pastures are suffering from. degradation. Overgrazing practised to maintain high stocking rates is regarded as one of the main causes. High stocking rates are deliberate and crucial decisions taken by the farmers, which appear paradoxical, even irrational given the state of knowledge regarding the consequences of overgrazing. The phenomenon however appears inextricably linked with the objectives that farmers hold. In this research those objectives were elicited first and from their ranking two, 'asset value of cattle (representing cattle ownership and 'present value of economic returns', were chosen to develop an original bi-criteria Compromise Programming model to test various hypotheses postulated to explain the overgrazing behaviour. As part of the model a pasture productivity index is derived to estimate the pasture recovery cost. Different scenarios based on farmers' attitudes towards overgrazing, pasture costs and capital availability were analysed. The results of the model runs show that benefits from holding more cattle can outweigh the increased pasture recovery and maintenance costs. This result undermines the hypothesis that farmers practise overgrazing because they are unaware or uncaring caring about overgrazing costs. An appropriate approach to the problem of pasture degradation requires information on the economics,and its interplay with farmers' objectives, for a wide range of pasture recovery and maintenance methods. Seen within the context of farmers' objectives, some level of overgrazing appears rational. Advocacy of the simple 'no overgrazing' rule is an insufficient strategy to maintain the long-term sustainability of the beef production systems in Central Brazil. (C) 2004 Elsevier Ltd. All rights reserved.