Purpose – To test the ability of major search engines, Google, Yahoo, MSN, and Ask, to distinguish between German and English-language documents Design/methodology/approach – 50 queries, using words common in German and in English, were posed to the engines. The advanced search option of language restriction was used, once in German and once in English. The first 20 results per engine in each l...

This paper attempts to deal with a ranking problem with a collection of financial reports. By using the text information in the reports, we apply learning-to-rank techniques to rank a set of companies to keep them in line with their relative risk levels. The experimental results show that our ranking approach significantly outperforms the regression-based one. Furthermore, our ranking models no...

This paper describes our proposed solution for the Yahoo! Learning to Rank challenge. The solution consists of an ensemble of three point-wise, two pair-wise and one list-wise approaches. In our experiments, the point-wise approaches are observed to outperform pairwise and list-wise ones in general, and the final ensemble is capable of further improving the performance over any single approach....

RÉSUMÉ. Optimiser le classement des résultats d’un moteur par un algorithme de learning to rank nécessite de connaître des jugements de pertinence entre requêtes et documents. Nous présentons les résultats d’une étude pilote sur la modélisation de la pertinence dans les moteurs de recherche géoréférencés. La particularité de ces moteurs est de présenter les résultats de recherche sous forme de ...

Canopy layer heat islands (CLHIs) in urban areas are a growing problem. In recent decades, the key issues have been how to monitor CLHIs at a large scale, and how to optimize the urban landscape to mitigate CLHIs. Taking the city of Wuhan as a case study, we examine the spatiotemporal trends of the CLHI along urban-rural gradients, including the intensity and footprint, based on satellite obser...

This paper describes the GIBIS team experience in the Predicting Media Interestingness Task at MediaEval 2017. In this task, the teams were required to develop an approach to predict whether images or videos are interesting or not. Our proposal relies on late fusion with rank aggregation methods for combining ranking models learned with different features and by different learning-to-rank algor...

The problem to replace a word with a synonym that fits well in its sentential context is known as the lexical substitution task. In this paper, we tackle this task as a supervised ranking problem. Given a dataset of target words, their sentential contexts and the potential substitutions for the target words, the goal is to train a model that accurately ranks the candidate substitutions based on...

Agricultural product review is playing increasing role in finance websites. Since manual text summarization needs large human efforts and also time consuming, we propose automatic summarization for agricultural product review in this paper. Firstly, we formulate the agricultural product as a quintuple, and present a framework to extract the features from agricultural product. Then we exploit le...

We propose a method for automatically labelling topics learned via LDA topic models. We generate our label candidate set from the top-ranking topic terms, titles of Wikipedia articles containing the top-ranking topic terms, and sub-phrases extracted from the Wikipedia article titles. We rank the label candidates using a combination of association measures and lexical features, optionally fed in...