Article review; Estimating Effectiveness of Twitter Messages with a Personalized Machine Learning

assume this is paper is a presentation, please do the following

use minimum (5 references).

• Identify the presenter, his/her affiliation and the context for the presentation.
• In 2-3 paragraphs, summarize the presentation: describe the focus and identify
the major points of the presentation. Do not insert your opinions in this part.
• If the presentation being critiqued is a research study, describe the type of
research, including purpose and methodology
• Comment on the presenter’s assumptions, methods and conclusions. What
was the presenter trying to accomplish? Did the presenter acknowledge and
respond to other points of view? How objective was the presenter? What new
ideas were presented? How do the presenter’s ideas compare with prevailing
views on the topic? What strengths or weaknesses did you notice in the
presenter’s methods and reporting?

Article to review

In Twitter, many aspects of retweeting behavior, which is the most effective indicator of spreading effectiveness of the tweet, have been researched, such as whether a reader will retweet a certain tweet or not. However, the total number of retweets of the tweet, which is the quantitative measure of quality, has not been well addressed by existing work. To estimate the number of retweets and associated factors, this paper proposes a procedure to develop a personalized model for one author. The training data comes from the author’s past tweets. We propose 3 types of new features based on the contents of the tweets: Entity, Pair, and Cluster features, and combine them with features used in prior work. The experiments on 7 authors demonstrate that comparing to the previous features only. Pair feature has a statistically significant improvement on the correlation coefficient between the prediction and the actual number of retweets. We studied all combinations of the 3

types of features, and the combination of the Pair and Cluster features has the best performance overall. As an application, this work can be used as a personalized tool for an author to evaluate his/her tweet before posting it, so that he/she can improve the tweet to achieve more attention.

Twitter as a platform of both the news media and social networks has been the subject of much research as of late. Most of the studies are interested in analyzing retweeting behavior, which is after a tweet posted by the author, some readers are attracted by the content of it and are willing to forward it and spread the information. The more the tweet is retweeted, the wider it spreads, so being retweeted shows how influential the tweet is.

Some work addressed questions like “What kind of author is more probably retweeted?” [13]. But the tweet author might be not helped by the answers because they are not constructive advice. It is true but useless to point out that to receive more retweets you need more followers, because the number of followers cannot be changed in a short time. The number of followers is the result of good tweets but not the other way around. Some other work answered the questions like “Which reader will retweet the tweet?” [8]. However, most people like to post tweets in public rather than only sending them to the specified readers.

So far as we know, the question that “How to anticipate the popularity of a tweet?” hasn’t been well addressed. The question is motivated by the observation that some tweets are more popular than the others, even they are from the same author. It’s a difficult problem to solve because it’s much harder than telling why a tweet from a celebrity is more

influential than from a regular person, or telling whether a football fan will be interested by the tweet or not. It’s a crucial problem because as an author what he / she really wants to know is “Can I write my tweet in a better way so that more people can see it?” So a procedure that addresses this could have substantial marketing or political value.

This work estimates how effective the tweet is in the aspect of the specified author, by analyzing the features from the author’s historical tweets, and training a personalized model for prediction. In additions to features which have proved to be important previously, we develop the additional Entity, Pair, and Cluster features to extract more abstract information from tweet, then train the machine learning regression models to predict the effectiveness of the tweet, which is the number of retweets.

iii