MAII FEI FAT and NAO TAI TAO ( 賣飛佛 與 鬧太套)

As time goes on, we find this course is becoming more and more challenging. Recently we are busy preparing for the final project, at the same time, learning programming with python.

Since this is a really multidiscipline course, this time, I’d love to talk about something more relaxing, the cyber-culture in social media. More specifically, some fun web slang!

Do you know this two handsome guys? Yes! They are Ng Cheuk Hei from HK and Huang Xiaoming from mainland, both are very popular actors. Two funny slang, “Mai Fei Fat” and “Not At All” were created because of them. 

In a TV commercial, Ng said “Tai Lin, my favorite”. With his strong HK accent, it just sounds like “Tai Lin, Maii Fei Fat”. Then the slang was created and used for mocking Ng and just for fun. Now not only young people in HK use this word, you can also see some figures in TV series using it.

Very similarly, netizens in mainland use the slang “Nao Tai Tao” to laugh at Huang’s terrible oral English.  Because he pronounced “Not at all” like “Nao Tai Tao”.

Actually, their English are not too bad. Maybe this two words became popular just because there are many people jealous of Ng and Huang’s handsome look. LOL

Young people always like kuso, suck gags, anti-humor and another funny but stupid things. That’s why GOLDEN Forum in HK as well as Tianya and Tieba in mainland could be so hot. I found that even though netizens in HK and mainland use different social platform, but their cyber-culture really have some similarities. 

The other interesting example is “MK Look” and “Sha Ma Te”. Both of these neologisms are used to describe certain dressing styles of some people, who want to be fashion, but could only afford to cheap, poor-quality and weird clothes. Those kind of clothes could be found in Mong Kok (a shopping area in HK) and Taobao (a very influential online shopping site operated by Alibaba). Behind “MK Look” and “Taobao Feel” are two subcultures originated from cyberspace, “Mong Kok Culture” and “Sha Ma Te Culture”.

Everyday, more and more new words would emerge online, some of them may even penetrate into everyone’s daily life.  Our course project targets social media, exactly where most of new words come from. When we analyze netizens’ comments, we have to figure out the meaning of the internet slangs they use. This may increase the difficulty of our natural language processing and data mining.  

Was That...Sarcasm?

In the past weeks, we discussed about Natural Language Processing (NLP) and Sentiment Analysis. And I learned that the goal of sentiment analysis is to make machines understand the opinions and attitudes behind texts written in natural language.

Then I raised a question. How about sarcasm?

We know that it is even difficult for human to understand ironies. If you are also a fan of the popular sitcom the Big Bang Theory (TBBT), you must know that in this series, Dr. Sheldon Cooper, with an IQ of 187, could hardly recognize sarcasms. His roommate Leonard even made a sarcasm sign to let he know what Penny said to him is sarcastic.

 Actually researchers for NLP are trying to solve this problem since 1990s, and they called this question irony detection.

If we are talking to a person face to face, we may detect irony from his or her facial expression or tone. However, detecting sarcasms from text and recognizing those sarcasms real polarities are absolutely more challenging.

Carvalho et al. (2009) suggest that emoticons, onomatopoeic expressions for laughter, heavy punctuation marks, quotation marks and positive interjections, could help computers  identify ironic sentences in user generate content (UGC).

While another group from Brazil  (Vanin et al. 2013) focus on Tweet, they summarize 15 expression patterns of irony from Portuguese tweets. They combine “Pattern detection” and “Manual tagging” to improve their system’s accuracy for irony detection. 


The results achieved with three classifiers Näıve Bayes (NB), support vector machines (SVM), and decision trees (DT), are satisfactory, both in terms of classification accuracy, as well as precision, recall, and F-measure.

The work on automatic irony processing is scare, a lot of effort still need to be done. How to make computer understand sarcasm? It is a interesting but tough question.

Reference:

CARVALHO, P., SARMENTO, L., SILVA, M.J. AND DE OLIVEIRA, E., 2009. Clues for 
                   Detecting Irony in User-generated Contents: Oh...‼ It’s “So Easy” ;-). In 
                   Proceedings of the 1st International CIKM Workshop on Topic-sentiment 
                   Analysis  for Mass Opinion. TSA ’09. New York, NY, USA: ACM, pp. 53–56.

REYES, A. AND ROSSO, P., 2011. Mining Subjective Knowledge from Customer Reviews: A 
             Specific Case of Irony Detection. In Proceedings of the 2Nd Workshop on 
             Computational Approaches to Subjectivity and Sentiment Analysis. WASSA ’11. 
             Stroudsburg, PA, USA: Association for Computational Linguistics, pp. 118–124.

VANIN, A.A., FREITAS, L.A., VIEIRA, R. AND BOCHERNITSAN, M., 2013. Some Clues on 
            Irony Detection in Tweets. In Proceedings of the 22Nd International Conference on 
           World Wide Web Companion. WWW ’13 Companion. Republic and Canton of Geneva, 
           Switzerland: International World Wide Web Conferences Steering Committee, pp.   
           635– 636.

print('Hello world')

 Hello, everyone. I'm Jiang Yi 蔣禕. And this is my new blog for the course IEMS5723.

 You can also find me at WeChat/QQ (804211369), Weibo, Facebook, Renren, Zhihu or Linkedin.

Social Media Analysis is a very interesting topic. As an active user, who spend(#or waste) at least one hour (#very  conservative estimation) on SNSs, sometimes I would like to review what I have done and learn about how I behavior at SNSs.

And I found some apps. The first one is “Mydata” on Weibo, which could tell me male-to-female ratio and the astrological sign distribution of my followers. (#OMG, so many  virgos!)

The second is “FrienDoc” on Renren, which tells me the users who “care” about me most, or interact with me most frequently.

The last one is developed by Natural Language Processing and Computational Social Science Lab, Tsinghua University, named “Key Words”. You can find this app in many different SNSs. It abstract the key words from the content I posted, sort them by how frequent those words appear, and visualize the result. The key word appears for the most of number would be in the middle of the picture and would have bigger size than any other keywords.  

(#the key words of my Douban)

All those three apps, however, are only focus on one user’s content and behavior. Analyzing social media would more complex and would get more valueable results.

This course also give me an opportunity as well as motivation to learn Python, the powerful coding language that we will use for this course’s projects. I have a good feeling of this language when I heard about it for the first time. It’s not because I am a geek or a coding lover, but I really like Monty Python’s comedy movies. (#  XD  )

I started to learn Python above one week ago. It’s seems that Python is not as complex as the languages I am more familiar with, such as C and Java. As far as I am concerned, by using Python, coders could spend more time and energy to solve the really problem, instead of solving syntax or data structure problems. Python would be a relatively mature language for social media analyzing. I hope I will have many interesting discovers by mining data with Python.

Now it’s 21 September 23:40. As the DDL is approaching, this would be the end of my first post. Next time I may do some more sericous work.

Thank you for reading. Thank you for your kindly comments. 多謝你哋！唔該晒！