2014年9月21日 星期日

Post 1 - Understanding what is "Social Media Analytics"

Basis ideas of Social Media Analytics

In lecture 1, we discussed what are social media and what is social media analytics.  
We used the definition of social media as a starting point to start the discussion of how we can analyse the complicated reaction within the social media.

Web 2.0 and User Generated Content (UGC) are two elements to drive Social Media. Social media is a global phenomenon, it enabled the information sharing between a groups and communities. People are enable express their views,interest, religion and opinions via the social media. Social media offers a place to let us to understand people and their behaviour, it also provided us a method to use the macro point of view to analyse global phenomenon.

Social media analytic is an engineering point of view to develop and evaluate the information tools and frameworks to extract, analyse, summarise and visualise information.  During the discussion of "Case Studies 3 - How can we improve face recognition using social network information", face recognition system is alway perform badly in different pose, illumination and expression but social media information such as friendship information can improve the face recognition algorithm. It mean that the system can "guess" the face more accurate by knowing the relationship.  Through this cases we can learn that the social media analytics data is not just a "output" it can also be a "input" to improve a real social media system.


3 則留言:

  1. 作者已經移除這則留言。

    回覆刪除
  2. 作者已經移除這則留言。

    回覆刪除
  3. It is really fantastic idea that you can relate face recognition technique with social media! The accuracy problem is what the face recognition is facing because few data are available for a person who just appear in front of the camera or his face cannot be reflected clearly on face recognition device. But with social connection data, the face recognition system can have a reasonable guess to narrow the range of possible names. Your idea is really good to improve the existing algorithms of face recognition!

    BTW, you are earnestly welcome to my blog for a visit if you are interested. Any comments and ideas are appreciated. Thanks for your sharing!

    回覆刪除