Results 1 -
8 of
8
Characterizing the Life Cycle of Online News Stories Using Social Media Reactions
"... This paper presents a study of the life cycle of news articles posted online. We consider user activity both from the perspective of their visitation patterns and from their social media reactions. We show that we can use this information to characterize distinct classes of articles, and that we can ..."
Abstract
-
Cited by 8 (3 self)
- Add to MetaCart
(Show Context)
This paper presents a study of the life cycle of news articles posted online. We consider user activity both from the perspective of their visitation patterns and from their social media reactions. We show that we can use this information to characterize distinct classes of articles, and that we can use social media reactions to predict future visitation patterns early and accurately. We validate our methods using qualitative analysis as well as quantitative analysis on data from the website of Al Jazeera in English, for a set of articles generating more than 3,000,000 visits and 200,000 social media reactions. We show that it is possible to predict the overall traffic an article will receive with the first ten minutes of social media reactions; the prediction accuracy is equivalent to the one based solely on visits after three hours. We also describe significant improvements on the accuracy of the prediction of shelf-life for news stories.
Predicting the Popularity of Web 2.0 Items Based on User Comments∗
"... In the current Web 2.0 era, the popularity of Web resources fluctu-ates ephemerally, based on trends and social interest. As a result, content-based relevance signals are insufficient to meet users ’ con-stantly evolving information needs in searching for Web 2.0 items. Incorporating future populari ..."
Abstract
-
Cited by 3 (2 self)
- Add to MetaCart
(Show Context)
In the current Web 2.0 era, the popularity of Web resources fluctu-ates ephemerally, based on trends and social interest. As a result, content-based relevance signals are insufficient to meet users ’ con-stantly evolving information needs in searching for Web 2.0 items. Incorporating future popularity into ranking is one way to counter this. However, predicting popularity as a third party (as in the case of general search engines) is difficult in practice, due to their lim-ited access to item view histories. To enable popularity prediction externally without excessive cra-wling, we propose an alternative solution by leveraging user com-ments, which are more accessible than view counts. Due to the sparsity of comments, traditional solutions that are solely based on view histories do not perform well. To deal with this sparsity, we mine comments to recover additional signal, such as social influ-ence. By modeling comments as a time-aware bipartite graph, we propose a regularization-based ranking algorithm that accounts for temporal, social influence and current popularity factors to predict the future popularity of items. Experimental results on three real-world datasets — crawled from YouTube, Flickr and Last.fm — show that our method consistently outperforms competitive base-lines in several evaluation tasks.
Predicting the Popularity of Online Serials with Autoregressive Models
"... Recent years have witnessed the rapid prevalence of online serials, which play an important role in our daily entertainment. A critical demand along this line is to predict the popularity of online seri-als, which can enable a wide range of applications, such as online advertising, and serial recomm ..."
Abstract
-
Cited by 1 (0 self)
- Add to MetaCart
(Show Context)
Recent years have witnessed the rapid prevalence of online serials, which play an important role in our daily entertainment. A critical demand along this line is to predict the popularity of online seri-als, which can enable a wide range of applications, such as online advertising, and serial recommendation. However, compared with traditional online media such as user-generated content (UGC), on-line serials have unique characteristics of sequence dependence, release date dependence as well as unsynchronized update regu-larity. Therefore, the popularity prediction for online serials is a nontrivial task and still under-addressed. To this end, in this pa-per we present a comprehensive study for predicting the popular-ity of online serials with autoregressive models. Specifically, we first introduce a straightforward yet effective Naive Autoregressive (NAR) model based on the correlations of serial episodes. Further-more, we develop a sophisticated model, namely Transfer Autore-gressive (TAR) model, to capture the dynamic behaviors of audi-ences, which can achieve better prediction performance than the NAR model. Indeed, the two models can reveal the popularity gen-eration from different perspectives. In addition, as a derivative of the TAR model, we also design a novel metric, namely favor, for evaluating the quality of online serials. Finally, extensive experi-ments on two real-world data sets clearly show that both models are effective and outperform baselines in terms of the popularity prediction for online serials. And the new metric performs better than other metrics for quality estimation.
Video Popularity Prediction by Sentiment Propagation via Implicit Network
"... Video popularity prediction plays a foundational role in many aspects of life, such as recommendation systems and invest-ment consulting. Because of its technological and economic importance, this problem has been extensively studied for years. However, four constraints have limited most related wor ..."
Abstract
- Add to MetaCart
(Show Context)
Video popularity prediction plays a foundational role in many aspects of life, such as recommendation systems and invest-ment consulting. Because of its technological and economic importance, this problem has been extensively studied for years. However, four constraints have limited most related works ’ usability. First, most feature oriented models are inadequate in the social media environment, because many videos are published with no specific content features, such as a strong cast or a famous script. Second, many studies assume that there is a linear correlation existing between view counts from early and later days, but this is not the case in every scenario. Third, numerous works just take view counts into consideration, but discount associated sen-timents. Nevertheless, it is the public opinions that directly drive a video’s final success/failure. Also, many related ap-proaches rely on a network topology, but such topologies are unavailable in many situations. Here, we propose a Dual Sentimental Hawkes Process (DSHP) to cope with all the problems above. DSHP’s innovations are reflected in three ways: (1) it breaks the ”Linear Correlation ” assumption, and implements Hawkes Process; (2) it reveals deeper fac-tors that affect a video’s popularity; and (3) it is topology free. We evaluate DSHP on four types of videos: Movies, TV Episodes, Music Videos, and Online News, and compare its performance against 6 widely used models, including Trans-lation Model, Multiple Linear Regression, KNN Regression, ARMA, Reinforced Poisson Process, and Univariate Hawkes Process. Our model outperforms all of the others, which in-dicates a promising application prospect.
New and Improved: Modeling Versions to Improve App Recommendation
"... Existing recommender systems usually model items as static — un-changing in attributes, description, and features. However, in do-mains such as mobile apps, a version update may provide substan-tial changes to an app as updates, reflected by an increment in its version number, may attract a consumer ..."
Abstract
- Add to MetaCart
(Show Context)
Existing recommender systems usually model items as static — un-changing in attributes, description, and features. However, in do-mains such as mobile apps, a version update may provide substan-tial changes to an app as updates, reflected by an increment in its version number, may attract a consumer’s interest for a previously unappealing version. Version descriptions constitute an important recommendation evidence source as well as a basis for understand-ing the rationale for a recommendation. We present a novel frame-work that incorporates features distilled from version descriptions into app recommendation. We use a semi-supervised topic model to construct a representation of an app’s version as a set of latent topics from version metadata and textual descriptions. We then dis-criminate the topics based on genre information and weight them on a per-user basis to generate a version-sensitive ranked list of apps for a target user. Incorporating our version features with state-of-the-art individual and hybrid recommendation techniques signif-icantly improves recommendation quality. An important advantage of our method is that it targets particular versions of apps, allowing previously disfavored apps to be recommended when user-relevant features are added.
Predicting Pinterest: Automating a Distributed Human Computation
"... Everyday, millions of users save content items for future use on sites like Pinterest, by “pinning ” them onto carefully categorised personal pinboards, thereby creating personal taxonomies of the Web. This paper seeks to understand Pinterest as a distributed hu-man computation that categorises imag ..."
Abstract
- Add to MetaCart
(Show Context)
Everyday, millions of users save content items for future use on sites like Pinterest, by “pinning ” them onto carefully categorised personal pinboards, thereby creating personal taxonomies of the Web. This paper seeks to understand Pinterest as a distributed hu-man computation that categorises images from around the Web. We show that despite being categorised onto personal pinboards by in-dividual actions, there is a generally a global agreement in implic-itly assigning images into a coarse-grained global taxonomy of 32 categories, and furthermore, users tend to specialise in a handful of categories. By exploiting these characteristics, and augmenting with image-related features drawn from a state-of-the-art deep con-volutional neural network, we develop a cascade of predictors that together automate a large fraction of Pinterest actions. Our end-to-end model is able to both predict whether a user will repin an image onto her own pinboard, and also which pinboard she might choose, with an accuracy of 0.69
Prediction of Popular Content from Social Media Mining
"... Abstract-In recent trends social media websites, such as ..."
1TrendLearner: Early Prediction of Popularity Trends of User Generated Content
"... Accurately predicting the popularity of user generated content (UGC) is very valuable to content providers, online advertisers, as well as social media and social network researchers. However, it is also a challenging task due to the plethora of factors that affect content popularity in social syste ..."
Abstract
- Add to MetaCart
Accurately predicting the popularity of user generated content (UGC) is very valuable to content providers, online advertisers, as well as social media and social network researchers. However, it is also a challenging task due to the plethora of factors that affect content popularity in social systems. We here focus on the problem of predicting the popularity trend of a piece of UGC (object) as early as possible, as a step towards building more accurate popularity prediction methods. Unlike previous work, we explicitly address the inherent tradeoff between prediction accuracy and remaining interest in the object after prediction, since, to be useful, accurate predictions should be made before interest has exhausted. Moreover, given the heterogeneity in popularity dynamics across objects in most UGC applications, this tradeoff has to be solved on a per-object basis, which makes the prediction task harder. We propose to tackle this problem with a novel two-step learning approach in which we: (1) extract popularity trends from previously uploaded objects, and then (2) predict trends for newly uploaded content. The first step exploits a time series clustering algorithm to represent each trend by a time series centroid. We propose to treat the second step as a classification problem. First, we extract a set of features of the target object corresponding to the distances of its early popularity curve to the previously identified centroids. We then combine these features with content features (e.g., incoming links, category), using them to train classifiers for prediction. Our experimental results for YouTube datasets show that we can achieve Micro and Macro F1 scores between 0.61 and 0.71 (a gain of up to 38 % when compared to alternative approaches), with up to 68 % of the views still remaining for 50 % or 21 % of the videos, depending on the dataset. We also show that our approach
