Short Text Mining: Machine Learning and Statistical Modelling Approaches Compared
Omar H. Al-Barahamtoshy · International Journal of Engineering & Technology · 2019
With the growth of technology, social media has gained popularity and now plays a key role in modern day to day communication. Given such trend, social media has gained increasing influence on our society to the extent it has become a part of our language to say I am going to “Tweet” about some thought. Like any community driven content, people find complex means to interact with each other. Twitter offers people the ability to tag their tweets with hashtags to specify the topic of tweet’s content. However, like any community driven convention, there exists many tweets which do not have hashtags. In this paper, we seek to explore the methods in the literature that can categorize tweets without hashtags. We have evaluated one method, which proved to be very promising due to its flexibility and extensibility to many applications. We also discuss future enhancement and extension possibilities, and provide a critique of the current method’s drawbacks.