Nuno Moniz

Resampling Approaches to Improve News Importance Prediction

The methods used to produce news rankings by recommender systems are not public and it is unclear if they reflect the real importance assigned by readers. We address the task of trying to forecast the number of times a news item will be tweeted, as a proxy for the importance assigned by its readers. We focus on methods for accurately forecasting which news will have a high number of tweets as these are the key for accurate recommendations. This type of news is rare and this creates difficulties to standard prediction methods. Recent research has shown that most models will fail on tasks where the goal is accuracy on a small sub-set of rare values of the target variable. In order to overcome this, resampling approaches with several methods for handling imbalanced regression tasks were tested in our domain. This paper describes and discusses the results of these experimental comparisons.

A Framework for Recommendation of Highly Popular News Lacking Social Feedback

Social media is rapidly becoming the main source of news consumption for users, raising significant challenges to news aggregation and recommendation tasks. One of these challenges concerns the recommendation of very recent news. To tackle this problem, approaches to the prediction of news popularity have been proposed. In this paper, we study the task of predicting news popularity upon their publication, when social feedback is unavailable or scarce, and to use such predictions to produce news rankings. Unlike previous work, we focus on accurately predicting highly popular news. Such cases are rare, causing known issues for standard prediction models and evaluation metrics. To overcome such issues we propose the use of resampling strategies to bias learners towards these rare cases of highly popular news, and a utility-based framework for evaluating their performance. An experimental evaluation is performed using real-world data to test our proposal in distinct scenarios. Results show that our proposed approaches improve the ability of predicting and recommending highly popular news upon publication, in comparison to previous work.

Time-Based Ensembles for Prediction of Rare Events in News Stream

Thousands of news are published everyday reporting worldwide events. Most of these news obtain a low level of popularity and only a small set of events become highly popular in social media platforms. Predicting rare cases of highly popular news is not a trivial task due to shortcomings of standard learning approaches and evaluation metrics. So far, the standard task of predicting the popularity of news items has been tackled by either of two distinct strategies related to the publication time of news. The first strategy, a priori, is focused on predicting the popularity of news upon their publication when related social feedback is unavailable. The second strategy, a posteriori, is focused on predicting the popularity of news using related social feedback. However, both strategies present shortcomings related to data availability and time of prediction. To overcome such shortcomings, we propose a hybrid strategy of time-based ensembles using models from both strategies. Using news data from Google News and popularity data from Twitter, we show that the proposed ensembles significantly improve the early and accurate prediction of rare cases of highly popular news.

Resampling Strategies for Imbalanced Time Series

Time series forecasting is a challenging task, where the non-stationary characteristics of the data portrays a hard setting for predictive tasks. A common issue is the imbalanced distribution of the target variable, where some intervals are very important to the user but severely underrepresented. Standard regression tools focus on the average behaviour of the data. However, the objective is the opposite in many forecasting tasks involving time series: predicting rare values. A common solution to forecasting tasks with imbalanced data is the use of resampling strategies, which operate on the learning data by changing its distribution in favor of a given bias. The objective of this paper is to provide solutions capable of significantly improving the predictive accuracy of rare cases in forecasting tasks using imbalanced time series data. We extend the application of resampling strategies to the time series context and introduce the concept of temporal and relevance bias in the case selection process of such strategies, presenting new proposals. We evaluate the results of standard regression tools and the use of resampling strategies, with and without bias over 24 time series data sets from 6 different sources. Results show a significant increase in predictive accuracy of rare cases associated with the use of resampling strategies, and the use of biased strategies further increases accuracy over the non-biased strategies.

Threshold-Bounded Influence Dominating Sets for Recommendations in Social Networks

The process of decision making in humans involves a combination of the genuine information held by the individual, and the external influence from their social network connections. This helps individuals to make decisions or adopt behaviors, opinions or products. In this work, we seek to investigate under which conditions and with what cost we can form neighborhoods of influence within a social network, in order to assist individuals with little or no prior genuine information through a two-phase recommendation process. Most of the existing approaches regard the problem of identifying influentials as a long-term, network diffusion process, where information cascading occurs in several rounds and has fixed number of influentials. In our approach we consider only one round of influence, which finds applications in settings where timely influence is vital. We tackle the problem by proposing a two-phase framework that aims at identifying influentials in the first phase and form influential neighborhoods to generate recommendations to users with no prior knowledge in the second phase. The difference of the proposed framework with most social recommender systems is that we need to generate recommendations including more than one item and in the absence of explicit ratings, solely relying on the social networks graph.

Empirical analysis of the Portuguese governments social network

The Portuguese governmental network comprising all the 776 ministers and junior ministers who were part of the 19 governments between the year 1976 and 2013 is presented and analyzed. The data contain information on connections concerning business and other types of organizations and, to our knowledge, there is no such extensive research in previous literature. Upon the presentation of the data, a social network analysis considering the temporal dimension is performed at three levels of granularity: network-level, subnetwork-level (political groups) and node-level. A discussion based on the results is presented. We conclude that although it fits two of the four preconditions of a small-world model, the Portuguese governmental network is not a small-world network, although presenting an evolution pointing toward becoming one. Also, we use a resilience test to study the evolution of the robustness of the Portuguese governmental network, pinpointing the moment when a set of members became structurally important.

The Utility Problem of Web Content Popularity Prediction

The ability to generate and share content on social media platforms has changed the Internet. With the growing rate of content generation, efforts have been directed at making sense of such data. One of the most researched problem concerns predicting web content popularity. We argue that the evolution of state-of-the-art approaches has been optimized towards improving the predictability of average behaviour of data: items with low levels of popularity. We demonstrate this effect using a utility-based framework for evaluating numerical web content popularity prediction tasks, focusing on highly popular items. Additionally, it is demonstrated that gains in predictive and ranking ability of such type of cases can be obtained via naïve approaches, based on strategies to tackle imbalanced domains learning tasks.

Resampling strategies for imbalanced time series forecasting

Time series forecasting is a challenging task, where the non-stationary characteristics of data portray a hard setting for predictive tasks. A common issue is the imbalanced distribution of the target variable, where some values are very important to the user but severely under-represented. Standard prediction tools focus on the average behaviour of the data. However, the objective is the opposite in many forecasting tasks involving time series: predicting rare values. A common solution to forecasting tasks with imbalanced data is the use of resampling strategies, which operate on the learning data by changing its distribution in favour of a given bias. The objective of this paper is to provide solutions capable of significantly improving the predictive accuracy on rare cases in forecasting tasks using imbalanced time series data. We extend the application of resampling strategies to the time series context and introduce the concept of temporal and relevance bias in the case selection process of such strategies, presenting new proposals. We evaluate the results of standard forecasting tools and the use of resampling strategies, with and without bias over 24 time series data sets from six different sources. Results show a significant increase in predictive accuracy on rare cases associated with using resampling strategies, and the use of biased strategies further increases accuracy over non-biased strategies.