دانلود رایگان ترجمه مقاله چالش های کاوش رسانه های اجتماعی – IEEE 2013

کاوش در رسانه های اجتماعی: چالش ها و فرصت ها

Mining Social Media: Challenges and Opportunities

چکیده :
فرصت های ایجاد شده با شبکه های اجتماعی منجر به هجوم میلیون ها کاربر به سایت هایی نظیر تویتر، فیسبوک و فوراسکوئر شده است. حتی سایت هایی نظیر آمازون به کاربران توانایی تعامل با یک دیگر را داده است، اگرچه این تعامل هدف اصلی بیان شده توسط این سایت نمی باشد. این سایت های شبکه اجتماعی و ویژگی های شبکه های اجتماعی تولید حجم زیادی از داده ها می کند به طوری که می توان از این داده ها برای نتیجه گیری در مورد رفتار اجتماعی مورد استفاده قرار داد که با استفاده از اندازه نمونه ساده مورد مطالعه قرار می گیرد. به این ترتیب امکان اعتبار سنجی و ارزیابی تئوری های اجتماعی را داده و تولید مدل های جدید برای شیوه تعامل افراد و گروه ها کرده و اهرمی را برای بهره گیری از قدرت جمعیت در اختیار می گذارد.

Abstract

The opportunities presented by social networking have led to millions of users flocking to sites like Facebook, Twitter, and Foursquare. Even sites like Amazon have added the ability for users to interact with one another, though it seems tangential to the site’s stated purpose. These social networking sites and social networking features generate massive amounts of data that can be used to draw conclusions about social behavior that could previously only be studied using relatively small sample sizes. This unlocks the ability to validate existing social theories, generate new models for how individuals and groups interact, and leverage the power of the crowd, among others.

I. INTRODUCTION

The early 21st century saw an explosion of online social networking activities with the advent of services like MySpace, Facebook, and Twitter. The massive popularity of these services lead to other social networking sites and services that leverage the power of social interaction. Foursquare and Facebook Places, along with some other small services, popularized adding geographic information to social media interactions. Foursquare users check in to locations, post tips for other users, and can earn badges. Foursquare also allows users to become ”Mayors” of locations by checking in to that location frequently, adding a competitive element that affects both a user’s friends and their non-friends. Even sites that are not traditionally social networking sites have added user interaction ability in the desire to take advantage of the popularity of social networking. For example, Amazon not only allows users to post reviews, but to evaluate those reviews for their helpfulness. This allows users to interact with each other, and provides Amazon with an automated method for discouraging fake, unhelpful reviews. Amazon also allows users to post their purchases and wish lists to Facebook and Twitter, adding another level of user interaction that bridges platforms. Just like other sites though, social networking sites also come and go. Nothing demonstrates this better than the migration en masse from MySpace to Facebook as Facebook opened up its availability to more users in the mid-2000s. This move prompted MySpace to launch a re-design of their website in January of this year. Obviously, the recency of the redesign mean that remains to be seen if this re-design is sufficient to attract users back to the site. Activity on social networks parallels activity in the real world, meaning that a user’s behavior online may be a useful indicator of their behavior in the real world, though certainly not every posts all of their activity online. Some users post almost none of their activity online. However, for those that do this means that, with enough data, we can ask many of the same questions and draw many of the same conclusions that social scientists ask and answer, but on a much larger scale. It is important to note that this does not invalidate the work of social scientists, only as a supplement. Social networks like Facebook and Twitter have an enormous number of users. According to newsroom.fb.com, Facebook has over a billion users, more than the population of the entire continent of Europe. If Facebook was its own country, it would be third largest in the world, behind only China and India. Twitter’s user count is similar in magnitude. According to analyst group Semiocast1, the service has over 500 million users. If Twitter was its own country it would also be the third largest. With user numbers in the hundreds of millions, it is inevitable that massive amounts of data will be generated. Twitter reported that 1.38 million posts (called ”Tweets”) occurred during the three hours of the State of the Union address this year2. On Election Day of 2012, users posted more than 31 million tweets about the election alone, reaching a maximum rate of more than 327,000 tweets per minute. The sheer amount of data involved makes thorough analysis of this data impossible using conventional techniques. To make sense of the massive amount of data generated by users on these services, new techniques are necessary to reduce the massive amount of information generated to a more manageable amount. To complicate the issue, much of social media data is noisy, making the discovery of meaningful signals in that data much more difficult. The techniques used to find information will naturally vary based on the objective of the analysis. We will use some illustrative examples to show some of the research issues tackled by the Data Mining and Machine Learning (DMML) Lab. We will use ongoing projects to demonstrate the methods, challenges, and opportunities involved. In this work, we focus on the following research issues: • Information Diffusion: Understanding the patterns that underly memes and virality in social media. • Privacy and Vulnerability: Understanding how user choices lead to leaking personal information. • Trust Prediction: Modeling the emerging and evolving way users develop trust relationships online. • Sentiment Analysis: Automatically extracting the emotional content of social media items. • User Migration: Understanding when and why users move from one service to another • Location-Based Social Networks: Learning about realworld behaviors through social media evidence. • Tools for Leveraging Social Media: Helping organizations make use of social media data through analysis.