PrePrint: Supporting Trust Assessment and Decision-Making in Coalitions
Modern multi-organisational coalitions are capable of bringing diverse sets of capabilities, assets and information sources to bear on complex and dynamic operations. However, the successful completion of these operations places demands on the trust between coalition partners. When it is necessary to rely on other partners, decision-makers must be able to make rapid and effective trust assessments and decisions. In this paper, we introduce computational trust mechanisms, and illustrate how they may be employed in coalition information acquisition. We discuss how these mechanisms can draw on different sources of evidence to make assessments of trust, and arrive at decisions about how to act when trust can be supplemented by controls. Finally, we discuss future directions for these systems, and highlight some challenges that remain.
PrePrint: Bird Flu Outbreak Prediction via Satellite Tracking
Advanced satellite tracking technologies have collected huge amounts of wild birds’ migration data. These data are very useful for biologists to understand birds’ dynamic migration patterns, to study correlations between the habitats, and to predict global spread trends of avian influenza. We transform the biological problem into a machine learning problem by converting the migratory paths of wild birds into graphs. Our first step of H5N1 outbreak prediction is to discover weighted closed cliques from the graphs by our mining algorithm HELEN (short for High-wEight cLosed cliquE miNing), which are then used by our learning algorithm HELEN-p to predict potential H5N1 outbreaks at habitats. We show that the prediction is more accurate in comparison with that by the traditional method on a migration data set obtained through a real satellite bird-tracking system. It is also confirmed by our empirical analysis that H5N1 spreads in a manner of high-weight closed cliques and frequent cliques.
PrePrint: Multimodal Sentiment Analysis of Spanish Online Videos
The number of videos available online and elsewhere is continuously growing, and with this the need for effective methods to process the vast amount of multimodal information shared through this media. This paper addresses the task of multimodal sentiment analysis, and presents a method that integrates linguistic, audio, and visual features for the purpose of identifying sentiment in online videos. We focus our experiments on a new dataset consisting of Spanish videos collected from the social media website YouTube and annotated for sentiment polarity. Through comparative experiments, we show that the joint use of visual, audio, and textual features greatly improves over the use of only one modality at a time. Moreover, we also test the portability of our multimodal method, and run evaluations on a second dataset of English videos.
PrePrint: MuSES: a Multilingual Sentiment Elicitation System for Social Media Data
We develop MuSES, a multilingual sentiment identification system, which implements three different sentiment identification algorithms. Our first algorithm augments previous compositional semantic rules by adding social media-specific rules. In the second algorithm, we define a scoring function to measure the degree of a sentiment, instead of simply classifying a sentiment into binary polarities. All such scores are calculated based on a large volume of customer reviews. Due to the special characteristics of social media texts, we propose a third algorithm, which takes emoticons, negation word position, and domain-specific words into account. In addition, we propose a label-free process to transfer multilingual sentiment knowledge between different languages. We conduct our experiments on user comments from Facebook, tweets from Twitter, and multilingual product reviews from Amazon. We publish MuSES datasets online.
PrePrint: A Language Model Approach for Retrieving Product Features and Opinions from Customer Reviews
In this paper, we introduce a new methodology for the retrieval of product features and opinions from a collection of free-text customer reviews about a product or service. The proposal relies on a language modeling framework that can be applied to reviews in any domain and language provided with a minimal knowledge source of sentiments or opinions (e.g., a seed set of opinion words). The methodology combines both a kernel- based model of opinion words (learned from the knowledge source of sentiments or opinions) and a statistical mapping between words to approximate a model of product features from which the retrieval is carried out. To validate the usefulness of the proposal, we carried out experiments over several collections of customer reviews about products from different industry domains and languages (specifically, English and Spanish). We also compare the obtained results on the retrieval of product features to closely related work on extracting product features from customer reviews.