Popularity Prediction Problem Research Articles

TV shows popularity prediction of genre-independent TV series through machine learning-based approaches

The use of social media has grown exponentially in recent years up to become a reflection of human social attitudes and to represent today the main channel for conducting discussions and sharing opinions. For this reason, the vast amount of information generated is often used for predicting outcomes of real-world events in different fields, including business, politics, and health, as well as in the entertainment industry. In this paper, we focus on how data from Twitter can be used to predict ratings of a large set of TV shows regardless of their specific genre. Given a show, the idea is to exploit features concerning the pre-release hype on Twitter for rating predictions. We propose a novel machine learning-based approach to the genre-independent TV show popularity prediction problem. We compared the performance of several well-known predictive methods, and as a result, we discovered that LSTM and Random Forest can predict the ratings in the USA entertainment market, with a low mean squared error of 0.058. Furthermore, we tested our model by using data of “never seen” shows, by deriving interesting results in terms of error rates. Finally, we compared performance against relevant solutions available in the literature, with discussions about challenges arousing from the analysis of shows in different languages.

Content Popularity Prediction Based on Quantized Federated Bayesian Learning in Fog Radio Access Networks

In this paper, we investigate the content popularity prediction problem in cache-enabled fog radio access networks (F-RANs). In order to predict the content popularity with high accuracy and low complexity, we propose a Gaussian process based regressor to model the content request pattern. Firstly, the relationship between content features and popularity is captured by our proposed model. Then, we utilize Bayesian learning to train the model parameters, which is robust to overfitting. However, Bayesian methods are usually unable to find a closed-form expression of the posterior distribution. To tackle this issue, we apply a stochastic variance reduced gradient Hamiltonian Monte Carlo (SVRG-HMC) method to approximate the posterior distribution. To utilize the computing resource of fog access points (F-APs) and also reduce the communication overhead, we propose a quantized federated learning (FL) framework combining with Bayesian learning. The proposed quantized federated Bayesian learning framework allows each F-AP to send gradients to the cloud server after quantizing and encoding. It can achieve a tradeoff between prediction accuracy and communication overhead effectively. Simulation results show that the performance of our proposed policy outperforms the considered baseline policies.

Co-Evolving Popularity Prediction in Temporal Bipartite Networks: A Heuristics Based Model

One of the big challenges of our modern life is to find the right items or contents on the Internet and particularly in social media. One way of addressing the information overload problem in social media is to predict the future trends and popularity of online items. The popularity of an item can be measured by its attractiveness, i.e., the number of times it is being used. This popularity prediction can be translated to a link prediction and ranking problem, which aims to predict the link gain of the items in a user-item interaction network. User-item interactions in an online environment can be modelled as a bipartite network, where a link represents an event, reflecting a user buys or collects an item. Popularity prediction problem in temporal bipartite networks is of great interest to researchers. In this study, we propose a heuristic based model which only consider nodes collective link gain in a recent past time window of time as well as total link gain. To evaluate our model’s efficiency, we tested them on co-evolving social media items. We also evaluated the models’ performance on five information retrieval metrics (i.e., Area Under the Receiver Operating Characteristic, Kendall’s rank correlation tau, Precision, Novelty, and temporal novelty). The proposed model does not need hyper-parameter learning, which makes it the best choice for highly temporal and data streaming scenarios.

Graph representation learning for popularity prediction problem: A survey

The online social platforms, like Twitter, Facebook, LinkedIn and WeChat, have grown really fast in last decade and have been one of the most effective platforms for people to communicate and share information with each other. Due to the word-of-mouth effects, information usually can spread rapidly on these social media platforms. Therefore, it is important to study the mechanisms driving the information diffusion and quantify the consequence of information spread. A lot of efforts have been focused on this problem to help us better understand and achieve higher performance in viral marketing and advertising. On the other hand, the development of neural networks has blossomed in the last few years, leading to a large number of graph representation learning (GRL) models. Compared with traditional models, GRL methods are often shown to be more effective. In this paper, we present a comprehensive review for recent works leveraging GRL methods for popularity prediction problem, and categorize related literatures into two big classes, according to their mainly used model and techniques: embedding-based methods and deep learning methods. Deep learning method is further classified into convolutional neural networks, graph convolutional networks, graph attention networks, graph neural networks, recurrent neural networks, and reinforcement learning. We compare the performance of these different models and discuss their strengths and limitations. Finally, we outline the challenges and future chances for popularity prediction problem.

Federated Learning-Based Content Popularity Prediction in Fog Radio Access Networks

In this paper, the content popularity prediction problem in fog radio access networks (F-RANs) is investigated. In order to obtain accurate prediction with low complexity, we propose a novel context-aware popularity prediction policy based on <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">federated learning</i> (FL). Firstly, user preference learning is applied by considering that users prefer to request the contents they are interested in. Then, users’ context information is utilized to cluster users efficiently by adaptive context space partitioning. After that, we formulate a popularity prediction optimization problem to learn the local model parameters by using the stochastic variance reduced gradient (SVRG) algorithm. Finally, FL based model integration is proposed to learn the global popularity prediction model based on local models using the distributed approximate Newton (DANE) algorithm with SVRG. Our proposed popularity prediction policy not only can predict content popularity accurately, but also can significantly reduce computational complexity. Moreover, we theoretically analyze the convergence bound of our proposed FL based model integration algorithm. Simulation results show that our proposed policy increases the cache hit rate by up to 21.5 % compared to existing policies.

Predicting the popularity of scientific publications by an age-based diffusion model

• This study focuses on the popularity prediction problem of scientific papers based on citation networks. • The proposed AD model is a mimic of the attention diffusion process along the citation networks. • The AD model can achieve better prediction accuracy than other benchmark methods. • The AD model can identify the future highly cited papers from massive numbers of new papers. Predicting the popularity of scientific publications has attracted much attention from various disciplines. In this paper, we focus on the popularity prediction problem of scientific papers, and propose an age-based diffusion (AD) model to identify papers that will receive more citations and be popular in the near future. The AD model mimics the attention diffusion process along the citation networks. An experimental study shows that the AD model can achieve better prediction accuracy than other benchmark methods. For some newly published papers that have not accumulated many citations but will be popular in the near future, the AD model can substantially improve their rankings. This improvement is critical, because identifying future highly cited papers from large numbers of new papers published each month would provide very valuable references for researchers.

Evolutionary Prediction of Nonstationary Event Popularity Dynamics of Weibo Social Network Using Time-Series Characteristics

A growing number of web users around the world have started to post their opinions on social media platforms and offer them for share. Building a highly scalable evolution prediction model by means of evolution trend volatility plays a significant role in the operations of enterprise marketing, public opinion supervision, personalized recommendation, and so forth. However, the historical patterns cannot cover the systematical time-series dynamic and volatility features in the prediction problems of a social network. This paper aims to investigate the popularity prediction problem from a time-series perspective utilizing dynamic linear models. First, the stationary and nonstationary time series of Weibo hot events are detected and transformed into time-dependent variables. Second, a systematic general popularity prediction model N- SEP 2 M is proposed to recognize and predict the nonstationary event propagation of a hot event on the Weibo social network. Third, the explanatory compensation variable social intensity (SI) is introduced to optimize the model N- SEP 2 M. Experiments on three Weibo hot events with different subject classifications show that our prediction approach is effective for the propagation of hot events with burst traffic.

Cross-Platform Event Popularity Analysis via Dynamic Time Warping and Neural Prediction

Nowadays, the primary media for information dissemination is shifting to online media. Events usually burst online through multiple modern online media. Therefore, predicting event popularity trends becomes crucial for online platforms to track pubic concerns and make appropriate decisions. However, few researches focus on events popularity prediction from a cross-platform perspective. Challenges origin from vast diversity from events and media, limited access to aligned datasets across different platforms and the great deal of noise in datasets. In this paper, we solve the cross-platform event popularity prediction problem by proposing a model named DancingLines, which is mainly composed of the following three parts. First, we propose TF-SW, a semantic-aware popularity quantification model based on Term Frequency with Semantic Weight, obtaining the event popularity based on Word2Vec and TextRank and generating Event Popularity Time Series(EPTS). Then, we propose DTW-CD, a pairwise time series alignment model derived from DTW with Compound Distance, aligning the EPTS on several platforms. Finally, we aggregate two time series and propose a neural based prediction model implementing Long Short-Term Memory with attention mechanism to obtain accurate predictions. Evaluation results based on large scale real-world datasets demonstrate that DancingLines can efficiently characterize, align, and predict event popularity on cross-platform.

Diversity of temporal influence in popularity prediction of scientific publications

Predicting the future influential papers is a challenging issue witch has attracted many attentions. In this paper, we focused on the temporal information of citations to study the popularity prediction problem from the perspective of citation dynamics. The experimental study of the APS citation data shows that the temporal decay rate of the influence of citations is decay with paper’s age, and the decay rate is a power-law distribution. We introduced the diversity of temporal decay rate of the influence of citations to predict the future popularity of papers, and proposed a diverse temporal decay method. The result shows that this method can improve the prediction accuracy compared with other popularity-based prediction methods. More importantly, this method can detect some of the newly published papers that haven’t accumulated many citations but will quickly become popular in the future.

GHSCN: A Graph Neural Network-Based API Popularity Prediction Method in Service Ecosystem

With the rapid development of technologies in the field of service computing, and increasing of complex business requirements, more and more large-scale service ecosystem emerges. Thus, many researches of service ecosystem focus on issues related to optimization such as service recommendation and load balancing, so the API popularity prediction problem studied in this paper, which is basis for this service ecosystem optimization, becomes a research hotspot in this field. However, many existing researches are predicting the popularity of APIs based on API functions, QoS, history usage patterns and social relationships, which are difficult to obtain and cannot reflect the overall structure of the underlying service ecosystem. Therefore, we propose an innovative API popularity prediction method in service ecosystem based on Graph Neural Network (GNN). Concretely, a Global-Service Ecosystem Network (GSEN) model is proposed firstly, for modeling a given service ecosystem to a network that can depict the complex structure of service ecosystem and the functions, QoS, history usage patterns and social relationships of APIs. Then, a Graph Heterogeneous Spatiotemporal Convolutional Network (GHSCN) model is proposed to predict the popularity of APIs based on GSEN, and for getting better prediction accuracy, four different Heterogeneous Spatiotemporal Convolutional Kernels are proposed to extract the features of different elements which have different mechanisms to affect the popularity of target API. Finally, extensive experiments based on the data crawled from ProgrammableWeb.com show that our method achieves a superior performance in API popularity prediction, and the importance of the introduction of our model to service ecosystems.

Joint User Access Mode Selection and Content Popularity Prediction in Non-Orthogonal Multiple Access-Based F-RANs

Non-orthogonal multiple access (NOMA) is regarded as a promising technology for the next-generation wireless communication system. Introducing NOMA into the fog radio access networks (F-RANs) is able to provide simultaneous transmissions to multiple users and significantly enhance F-RAN performance. However, due to the increasing number of users and the constraint of caching storage capacity, there exists a tradeoff between NOMA transmission performance and fronthaul saving. In this paper, a hierarchical game framework is presented to solve the joint optimization problem of user access mode selection and content popularity prediction in NOMA based F-RANs. More specifically, the access mode selection problem is formulated as an evolutionary game. The proposals’ evolutionary payoff expressions are derived by stochastic geometry tool, and the cost functions are related to the fog access point (F-AP) content placement profile as well as the fronthaul constraint. Moreover, the problem of what contents the F-AP should cache is modeled as a content popularity prediction problem, and based on both local and global user request states, a machine learning algorithm is presented to solve it. Simulation results validate the accuracy of analytical results and demonstrate our proposed algorithms can further improve the performance of NOMA based F-RANs.

Temporal popularity prediction of locations for geographical placement of retail stores

Modern retail chains, such as Starbucks and McDonald’s, seek for geographical locations that have higher possibility to bring the maximum profit to establish and open their new stores. For a retail store, the common indicator of profit is its popularity which can be measured by the number of people who had ever consumed there. If one can know the potential popularity of locations along the time or within a certain time period (e.g., Christmas holiday or some memorial days) in advance, it would facilitate the determination of locations that are more profitable for the geographical placement of new retail stores. In this paper, given a targeted retail chain as well as some candidate locations, we propose to predict the popularity values of these locations over time to support their decisions on placing new stores. We devise a semi-supervised learning model, Affinity-based Popularity Inference, to tackle the Temporal Location Popularity Prediction problem. The basic idea is that locations that share similar properties of geo-spatial venue characteristics and human mobility tend to have closer popularity values. Experiments conducted on Foursquare check-in data in New York City exhibit promising results on predicting the temporal popularity of retail chains Starbucks, McDonald’s, and Dunkin’ Donuts, comparing to interpolation, regression, and state-of-the-art supervised learning methods.

Predicting the implicit and the explicit video popularity in a User Generated Content site with enhanced social features

User Generated Content (UGC) sites like YouTube are nowadays entertaining over a billion people. Identifying popular contents is essential for these giant UGC sites as they allow users to request contents from a potentially unlimited selection in an asynchronous fashion. In this work, we conduct an analysis on the popularity prediction problem in UGC sites and complement previous work with two new aspects, namely differentiating contents that attract a lot of attention and that users really appreciate, and leveraging built-in social features to predict the content popularity immediately upon publication.To this end, we conduct an extensive measurement and analysis of BiliBili, a YouTube-like UGC site with enhanced social features including user following, chat replay, and virtual money donation. Based on datasets that contain over 2 million videos and over 28 million users, we characterize the video repository and the user activities, we analyze the video popularities, we propose graph models that reveal user relationships and high-level social structures, and we successfully apply our findings to build machine-learned classifiers to identify popular videos.

On the Prediction of Flickr Image Popularity by Analyzing Heterogeneous Social Sensory Data.

The increase in the popularity of social media has shattered the gap between the physical and virtual worlds. The content generated by people or social sensors on social media provides information about users and their living surroundings, which allows us to access a user’s preferences, opinions, and interactions. This provides an opportunity for us to understand human behavior and enhance the services provided for both the real and virtual worlds. In this paper, we will focus on the popularity prediction of social images on Flickr, a popular social photo-sharing site, and promote the research on utilizing social sensory data in the context of assisting people to improve their life on the Web. Social data are different from the data collected from physical sensors; in the fact that they exhibit special characteristics that pose new challenges. In addition to their huge quantity, social data are noisy, unstructured, and heterogeneous. Moreover, they involve human semantics and contextual data that require analysis and interpretation based on human behavior. Accordingly, we address the problem of popularity prediction for an image by exploiting three main factors that are important for making an image popular. In particular, we investigate the impact of the image’s visual content, where the semantic and sentiment information extracted from the image show an impact on its popularity, as well as the textual information associated with the image, which has a fundamental role in boosting the visibility of the image in the keyword search results. Additionally, we explore social context, such as an image owner’s popularity and how it positively influences the image popularity. With a comprehensive study on the effect of the three aspects, we further propose to jointly consider the heterogeneous social sensory data. Experimental results obtained from real-world data demonstrate that the three factors utilized complement each other in obtaining promising results in the prediction of image popularity on social photo-sharing site.

A Peek Into the Future: Predicting the Popularity of Online Videos

Predicting the popularity of online videos is an important task for the service design, advertisement placement, network management, and so on. In this paper, we tackle the challenge head-on by casting the popularity prediction problem into two consecutive tasks: online video future popularity level prediction and online video future view count prediction. We first predict the future popularity levels of online videos, based on a rich set of features and effective classification technique. Then, according to the popularity level transitions, we build specialized regression models to predict the future view count values. We validate our approach on the exhaustive dataset of a leading online video service provider in China, namely, Youku. The experimental results show that comparing with two state-of-the-art baseline models, our proposed method can significantly decrease the relative prediction errors of 32.25% and 19.82%, respectively. At last, we also discuss the model setup and feature importance of our method. We believe our work can provide direct help in practical for the interested parties of online video service, such as service providers, online advisers, and network operators.