Function words are words such as the which have a particular grammatical role

Function words are words such as the which have a

This preview shows page 3 - 5 out of 23 pages.

sentence, or specify the attitude or mood of the speaker. Function words are words, such as ‘‘the’’, which have a particular grammatical role but little identifiable meaning (Klammer et al. 2000 ). On the other hand, content words typically carry semantic content, bearing reference to the world independently of its use within a particular sentence. A word to which an independent meaning can be given, by reference to a world outside a sentence, in which the word may occur (Winkler 2012 ). We approach the problem of gender classification of tweets in a way similar to that used in classification of regular text by Argamon et al. ( 2003 , 2007 ). We extract features based on function words, part of speech n-gram tags and the most popular content words. From these features we expect to get a greater insight into the subtle differences in the types of linguistic features used by the two genders on social media. The most frequent function words and content words, as well as the part of speech n-grams used by either gender in non-formal communication, could be studied to generate new business, research, and social insights. In the extant research literature we have not come across any work where, in the gender classification of microblogs, function words and part of speech n-grams have been used as features. Our work adds to the existing body of knowledge in the following three ways: 1. As a first in the field, function words and part of speech n-grams have been used as features to classify microblog text. It’s a unique effort in the area to establish features that appear innocuous but could be relevant in classification of unstructured data in case of other related classification problems in social media. 2. We improve the classification accuracy by 7 % over the existing gender classification software with the above-mentioned features for a small dataset (3000 tweets). It logically follows that on increasing the dataset size the classification accuracy should further improve. However, we also need to ensure that the improvement achieved should not be at the cost of precision and Gender classification of microblog text based on authorial style 119 123
Image of page 3
recall of the classification job, thus we have incorporated F-measure. F-measure is a well-accepted measurement criteria in data mining research as it does not get affected by class imbalance and provides the harmonic mean for precision and recall. 3. Since, we have focused on using features that are independent of the text, the obtained results are more applicable universally than the results obtained by capturing text based features. A comparative study of the performance of the various features captured in the research has also been done. 2 Related work Regular text classification for authorship profiling has been addressed as a research problem since early 2000’s (Koppel 2002 ; Argamon et al. 2003 ). These authors classified formal textbooks based on writing style. They used British National Corpus (BNC) tagged corpora for training features for classification. This was one of the first works in the classification of authorship of formal text based on gender.
Image of page 4
Image of page 5

You've reached the end of your free preview.

Want to read all 23 pages?

  • Winter '18
  • Amrita Chakraborty
  • Naive Bayes classifier, Document classification, P. K. Bala

What students are saying

  • Left Quote Icon

    As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

    Student Picture

    Kiran Temple University Fox School of Business ‘17, Course Hero Intern

  • Left Quote Icon

    I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

    Student Picture

    Dana University of Pennsylvania ‘17, Course Hero Intern

  • Left Quote Icon

    The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

    Student Picture

    Jill Tulane University ‘16, Course Hero Intern

Stuck? We have tutors online 24/7 who can help you get unstuck.
A+ icon
Ask Expert Tutors You can ask You can ask ( soon) You can ask (will expire )
Answers in as fast as 15 minutes
A+ icon
Ask Expert Tutors