The video industry's big data mining is less than 1%.

In the big data age, the data is changing our thinking. There are a few pictures to let us see, the first picture, published in July 2013, a number of Chinese Internet, the second is the proportion of Chinese mobile internet users, I think we should pay attention to the growth rate of mobile phone is very high. At the same time to share with you a data, Archie in the mobile end of the layout is relatively early, we started in 2011 to predict when the mobile end of the flow will exceed the PC side. We originally had a forecast is at the end of 2013, but after the actual operation, in fact, in 2013 September 2 that day the mobile end of the flow is completely over the PC side, while the mobile end of the growth rate is much faster than the PC, so the future mobile terminal potential is very large.

The most important value of large data is prediction

For mobile end we want to see what is changing our mobile phone now, I think everyone here can think about, now we look at the frequency of mobile phones, I basically 5 minutes will certainly look at the phone. Because of your fragmentation time, and the frequency of the use of mobile phones, now if the phone is only pure phone call this tool has been very weak, we more of the machine as a source of access to information.

What is the amount of data generated by the Internet on a daily basis? I have a set of data, we can send 294 billion emails each day, can produce 34 billion pieces of information on the Twitter, can clone 168 million DVD. All of this shows that the data now produced is very large. In these huge data how do we dig, what this data can do for us. These data are now characterized by a very large scale, faster processing speed and a very wide variety of data. The big Data age is not a requirement now we are, as before, in the age of sampling samples, we may be very concerned about the accuracy of the data, and in the big data age I think the accuracy of the data can be ignored. Because the trend and trends observed from large data are the values that we need to study what the entire data represents. So for the analysis of the data, finally we hope to produce high value of the prediction, the largest of the most important one value is the prediction.

Large data makes user services more accurate

For Archie and PPS, (Archie 2013 years of April acquisition of PPS), on this platform, this (slide show) is just some basic data, on the PC side generated user 350 million, mobile end user 200 million, our entire 1.2 billion hours of content, 200 million registered users and 133 million-hour mobile end playback. What can big data do?

For the traditional industry, Wal-Mart is the earliest use of large data to do business forecasts, and is very good. Wal-Mart will link up some of the data that is fed back to Facebook and Twitter. Hurricanes often occur in the United States, and when they do, they will find that sales are going up when they are put together from sales figures, and that Wal-Mart's operations are piled up from big data mining.

And on the Internet, the most profound research on data should be Google. For the flu prediction, Google has found 83 accurate words from thousands of words to match the flu in terms of the data it searches for and the words it uses to search the data. They were able to predict where the flu would occur, and it was even quicker and more accurate than the U.S. health watchdog.

For the data, we ultimately hope that it will be able to play an important role in commercial applications, a very simple example, this is a 2013 Taiwan drama called "Two Dads." The traditional piece will tell us when it pushes the play, the play user is about 24-35 years old, and through Baidu User's search and Archie art play, we are accurate to the user playback, can be accurate to his recent search of what things, on what site, he recently have what needs, through these data analysis, I gave him more accurate and precise marketing. For large data on the Internet, in fact, from the original label to the crowd into the label of the individual, for the user to provide personalized services for users more accurate and more valuable, the value of our media will be reflected.

Around us there are ipad, PC including now the video broadcast in taxis, the future includes watches, microwave ovens, refrigerators and so on will have a screen, and the screen processing speed will be faster. At this time we have to consider, since the user is surrounded by so many screens, what content we can occupy the user more time, this is the core value of the media. There is a different screen behavior data, in the commute, the mobile end must be the most core, the most convenient users to obtain information and access to video tools. This is the mobile replaces the PC becomes the starting point of the internet every day, many users are waking up in the morning, may first take out the phone to see, on the micro-letter, micro-blog, and then see what happened, and finally quickly look, or throw to the room to sleep. Most people's habits may be more and more in the use of mobile terminals to obtain information, and the PC is more a work tool.

Video Opportunities under Large data

Let me say a few things about video that we might have a chance of coming up with big data. I'm sure we're familiar with this card house. "Solitaire" is the 2013 American drama, in the creation of the American drama, through the analysis of 30 million users of the message and analysis, the first theme selected by the political background, because he thought that paid users of the subject will be very interested. Then its director and actor also through data mining and data analysis, it is a Kevinspethy, Fincher, the BBC produced three elements in one television program, which may have better ratings.

and the TV series marketing, planning and promotion, including online after the analysis of the user is very extreme, it is mainly to use the data, these data marketing to do better digging. The card house, though much of the hype is about using big data mining, I don't think so. The show's success is manifold, its marketing tools, including the big-name Hollywood directors it used to make, and Kevin was replaced by other directors after the first three episodes.

We can understand that, in the current era of the Internet so advanced, everything can be turned into data, when the text when the creative these become data, these data is actually for the creation of our content, production has a security in the early stage, this than we do not have any data on the premise to do the creation, There may be more value in this data base.

I am also here to say that Archie Arts in 3721.html ">2014 year will also vigorously promote our homemade dramas, including self-made variety show." Last year, Archie's character heroes and now hit the "Life needs to debunk", for these plays in the entire production, we have done a lot of data analysis and data mining. 2nd, in the video website, its biggest characteristic is has the massive information. Users have two channels to find their content, these two channels is a search, search habits I think we all like, browsing information is too much, the other is recommended, for the recommendation there are many kinds of recommendations, which I think the use of large data to do a lot of, do the most extreme personalized recommendations, is to turn the user's behavior into data. We are now every day on the Archie and PPS on the production of data is very large, these data we dug even less than 1%, we need to do more digging. For the user behavior of this, such as what type of user to see what type of film, or the user labeled, I and you are a kind of people, in the personalized time, we see each other to do promotion, you like this content opportunities will be high. When all this becomes possible to record, analyze, and mine data, it actually brings a lot of convenience to the user. Users have a habit of searching for content, and they look at behavioral analysis and we offer it to him.

In fact there are a lot of data, and ultimately we want all the behavioral analysis of data to fall into business. The behavior data that users generate on this platform, communication, will become the basis of our analysis. When conducting data analysis, we can classify these people in a scientific way, then divide them into different groups, and then have different individuals within each group. For the user's sex, age, preferences and his intentions, will become a data, he is what kind of users, he likes to see what kind of movies and so on.

The huge user each has the different label attribute, regardless of me in the content application is good, or is doing the commercial excavation or not, compared with the customer group, we may have more value to excavate. Let me give you two examples one is "a search hundred should", behind is Baidu, we and Baidu search, the entire Baidu search user data through, when the user in Baidu search, he to love after the odd art, the search behavior will spread to Archie Art, to achieve a way of communication. His target crowd searched these things to provide him with some of the content that he is interested in, whether it is business content or other content, our entire data is based on the data of Baidu and Archie art overall data to do analysis.

We also have a project is "Dandelion" project, for Baidu users of hundreds of millions of times a day search, especially after the search, for a lot of content may be very hot content, such as ISO, South Korea's star many people in search. Now hit the TV series in search, there are many searches, many did not hit, but the customer search, I hope to have a precise solution, for this we have "dandelion" project, will provide accurate content.

About the future of TV Internet, in fact, the Internet block is blocked, we have to use a kind of mentality, to embrace this new change, because the big era is not the Internet, belongs to all mankind.