Along with the advance of the society, we are now living in a world of numerous messages. Everybody uses social platform to get and share informations related with him. More and more netizens have been used to social software, such as Facebook, Google+, Wechat and Qzone. We can acquire great data from these social circles, which will make a amazing contribution to learning more about users' thoughts and improving our service.
However，how to deal with such an enormous data resource? As we know, text content is easier to analyse. Here natural language processing is required. It is a field that involves computer science, artificial intelligence, linguistics, human-computer interface and so on. Through taking NLP, we can learn customs' ideas, sentiment and behavior. To complete what mentioned above, we also have to dispose the useless words and punctuation.
In order to compare the documents, we are to classify the text, using vector to estimate terms and method of probability and mathematical statistics. In addition, clustering can finish it without labelled data. During learning K-Means clustering algorithm, I thought of the knowledge of pattern recognition. They apply the similar method to divide the items into any clusters. And they are both unsupervised, in relation to artificial intelligence. The second figure shows us how the pattern recognition works. Though the objects they dispose are different，they both reflect the application of function and the beauty of mathematics.
Social Media Analytics is a advanced and practical course. I do believe I will enjoy the course and reap no little benefit.
Hello, young artist! First of all, I have to say that I also enjoy this course and gain a lot, and we can share some opinions with each other after class if possible. Having read your blog, I realize that you really have a clear idea about text content analysis. I believe it's crucial to understand it because the text contents which directly generated by users are the response of their emotion. By mining and analyzing that can help us know more about them. In addition, could you please tell me what's the exact meaning of your funny second figure? Thank you~
回复删除Hello! Due to Pattern Recognition I mentioned above, I used the figure to show what it is to individuals who hasnot learnt about that. If you have interest in the field of recognition or you are willing to share your opinion on social media with me, we can have s further communication.^_^
删除Gooooooooood. Seem you learned very well in class. Well, classifying the test is crucial which can be easier to analyze the content. Also K-Means clustering is a great algorithm which is different from classification. The process of classification is called supervised learning while clustering is unsupervised learning which means clustering needn't learn from data it got previous. Hope can discuss more with me. Tete, come to comment my blog. (hehaohui-iems5723.blogspot.hk)
回复删除what you said just correspond to our teacher, and I found a lot of new things in your blog, which is very good.
回复删除Cool! you blog introduce the way of data mining clearly. That is very useful.
回复删除So the text is about the method of dealing with data....I remembered that the teacher also talks some ways. K-means is also a way of those, but I'm not quite understand it. Can you help me with this? Thanks a lot.
回复删除Good! You express your personal understanding about social media. For the question of dealing with enormous data resource, you understand what our teacher said in class well. I'm still confused about it. Maybe I need your help.^_^
回复删除interesing post! I am extremly interested in your comparison between the algorithm we leart from the lecture, and pattern recognition, the pics your provided are not quite clear, could we discuss this further in private?
回复删除