The secret charm of big data is that it knows you better than you do. Do you think that every time you press the phone button, the action is the same? Haha, the pattern of Tucson has been broken.
Yang Zhen, vice president of technology from today's headlines, told children's shoes that the "black technology" they are testing can only infer your mood at that time from the time you clicked the button and the area of your finger. Your carelessness, anger or moving can be the basis for what kind of news is pushed to you in the background. In the future, the following scenarios can be expected:
If you are in the sadness of being dumped by the goddess, maybe the client will push for you-one hundred benefits of homosexuality. If you are in the joy of receiving this month's salary, maybe the customer will give you a push-when can I buy a toilet for 20 thousand a month in Beijing?
So what does this "just right" emotional balance have to do with big data? In fact, the speculation about your mood is based on the record of your normal clicks many times. This kind of behavior data "betrays" your emotions even before you realize it.
Yang Zhenyuan, vice president of headline technology today, is analyzing the average touch time of a key.
Your "posture" is really big data.
The statistics of bank daily trading accounts are not big data. The real behavior is how long each user waited after getting the number, how many users scolded their mothers, and how many users left too anxious and angry.
Yang Zhenyuan gave another example of today's headline application.
In fact, how long you stayed in an article, how far you rowed, how long you stayed in a new position, whether you read the comments or not, and how many comments you read can all be recorded in order. The next step is to evaluate readers' interests through algorithms.
Jiang Tao, the founder of CSDN, also mentioned that Wish, an American e-commerce platform, is using big data to recommend products that you might like "by looking at people's dishes" according to different data of each person. In three years, it has developed into one of the largest e-commerce companies in North America.
Therefore, the sad news is that if you want to hide your identity in the future, you should not only disguise yourself, forge your fingerprints, but even change your habit of clicking your mobile phone and reading articles.
Big data is: no one can be less.
If you want to know what percentage of people like GV, you just need to do a sample survey, and there is no need to investigate everyone. However, if you want to sell smart hardware products around the house, you need to check everyone's "unique" hobbies one by one.
All the data are indispensable, which is called "full processing". The producers of these data are the profit sources of major manufacturers.
Liu Peng, the chief architect of 360 commercial products, is a network celebrity. He emphasized on many occasions that full-scale processing is big data. He said that when it comes to personalized recommendation, computing advertisements and personal credit information, large-scale computing is inevitable.
From a technical point of view, the reason why big data can be so accurate is mainly due to technological progress. Sensing devices are widely used in all kinds of hardware, so that data that could not be recorded before can be recorded now.
Big data should not be used by people.
Big data should be given to machine decision-making, not to people.
This assertion, which is full of deep distrust of human beings, also comes from Liu Peng. In his eyes, big data is the food of the machine. There are basically only two kinds of human beings who can control big data: data scientists and statisticians.
IT enterprises have no possibility of raising a group of scientists. And people's judgment is often based on macro and strategy, and there can be no energy that varies from person to person. In contrast, machines judge more carefully than humans. For example, for each user, such as portraits and labels. Therefore, in order to make full use of big data, stupid humans should step aside for a while.
"A little wrong" big data is better to use.
The word "data" naturally gives people a perfect and accurate feeling. In this regard, big data should challenge your bottom line. As a big bull in the field of digital advertising, Liu Peng emphasized that big data can have the ambiguous attribute of semi-consistency. In other words, data errors and losses are allowed.
Nani? Is wrong data also good data? That's right. Because of the huge amount of data and low value density, analysts often need to select some characteristic data for processing, which may be simplified before processing. So the final result of big data is hard to be confused, but it hits the nail on the head.
So, if someone recommends pomegranate to you from Love People's Network, don't get angry first. You may just be a mistake in big data.
Insurance companies like being with big data best.
If you are a reckless person, it is undoubtedly your auto insurance company that wants to know this situation most, and your premium will definitely remain high; If you are a cautious person, what you want to know most is the insurance company, because it can attract you to take out insurance with discounted premium.
You even have an accurate "risk rate" figure on you. This frightening figure is precisely the source of insurance companies' profits. Because individuals without such big data cannot calculate their own risk rate. Insurance companies take advantage of this information asymmetry, providing one thousandth of the insured price to a person with one thousandth of the risk rate, which is equivalent to earning ten times the profit.
Privacy depends on technological progress.
The data is more sinister than it looks, which is the general consensus of people in the big data industry. Even if you hide sensitive information such as your name and telephone number, only the records of your contact with others are kept, and people who are familiar with you can completely guess your identity. At present, the security of big data, under the malice of others, seems inadequate.
Privacy issues, the system can only solve 20 points, and the remaining 80 points must be solved by technological progress.
Liu Peng said. It seems hopeless to expect the market to go back to the pre-big data era.
How to accurately count how many people like Aoi sora and how many people like it, without revealing who likes Mr. Cang and who likes Mr. Wu, is the frontier research of big data at present.
No matter how strict the big data policy is, there is no reliable security technology, and the security of data is impossible to talk about. Privacy algorithm, data desensitization, data isolation. Are the direction of research. Before that, your big data was still in a relatively dangerous state. This is also why the current law does not prohibit the sale of data, but the giants are afraid to sell data. ? Of course, the relatively low market price of large databases is also an important reason.
China SDCC Software Developers Conference. Founded in 2007 by CSDN, the world's largest Chinese IT community, once a year. The theme is the next generation software development technology trend and its profound influence on various industries, and it is famous for its dry goods.
The above are some cold knowledge about big data that you may not know. For more information, you can pay attention to Global Ivy and share more dry goods.