Today's headline Zhang Yixi: Big Data is "God" in cyberspace

Source: Internet
Author: User
Keywords Today's headlines algorithms Zhang Yixi
Tags analysis based behavior big data client clients company customized

This is an era of information explosion, so the message filtering and personalized push based on large data, it becomes a very important thing. And that is why today's headlines can stand out from many news clients.

August 2012, today's headline online, with its unique large data analysis and customized recommendations and quickly received hot. From the information client to the recommendation engine for data mining, which is less than three years old, the company's valuations have reached $500 million trillion, and these are the value of the big data algorithms that are behind today's headlines.

How do you infer your preferences? Better algorithm than big data

When you travel in the field, open today's headlines, you will suddenly find that it for you to push the message, more than a few local travel strategy. When you can't sleep at night, when you open the app, the first thing you see is emotional content. These are one of the scenes that are used by today's headline users, and you may wonder, how does it make you better? Today's headline, the VP of Technology for Algorithmic data, gave us the answer.

"The personalized recommendation algorithm is based on voting," Yang said, and the core idea of today's headline-personalized recommendation is actually voting. "In fact, personalized recommendation is not a machine for you to recommend, but a person to recommend each other." ”

He compares the personalization algorithm to the method of counting beans, each person take a bean, like which article will give the beans to this article, and then by the machine to count, and finally get the result. According to this algorithm, the same two articles, the current three people are the first article to make "vote", then, to the fourth person pushed the inevitable is the first article.

In addition, the crowd can be based on age, gender, geography, occupation and other classifications, and the article can be keywords, release time, the region and other categories, and then combined with user feedback data, just "you care, is the headline." Yang Zhenguo said that today's headline recommended method is also machine learning algorithm, in the mobile internet era there are many beans produced, when you take a cell phone each read an article, your every click and search, all kinds of behavior, will be recorded to produce data.

As Yang originally said:

"More data beats decoupled algorithms, large numbers are better than good algorithms, the arrival of the big Data era really released the value and significance of these algorithms." ”

The next five years of headline today, machines can erase the information gap

After three years of entrepreneurship, Zhang Yixi's team has accumulated 220 million user habits and data. How can this data be better to play? So they had a press conference on "arithmetic", and for the first time, they showed the world what a group of users today's headlines were.

According to their data reports, male users accounted for 65% of the population, while women had fewer users, but their average stay was 8 minutes, significantly higher than 5.5 minutes for men.

"Guangxi people love the best, Hebei people love to see the jokes, the most concerned about the political affairs of Shanxi people, the most concerned about the gossip is Tianjin." ”

This is the headline of today's headlines based on user habits. Zhang Yixi explained that

"The past is to rely on people to sum up knowledge, now through the system, learning the user's behavior characteristics to store wisdom." A lot of people ask me, what will the headlines be like in the next 5 years? In fact, people look at my research app, probably know what I'm focusing on. ”

Zhang Yixi said while the mobile phone more than 400 apps show everyone. He himself summed up the direction of three efforts:

One is to let more kinds of information, including text, pictures, short video How to give users better in more scenarios. Second, we can grasp the user more and more habits and data, so that users based on the same interests to better communicate and interact. Third, we hope to connect more O2O services so that today's headlines can become "Life today".

So what about the algorithms and big data behind today's headlines?

Zhang Yixi that in the next five years, more and more people will be interested in, behavior is projected into the online world, when the data more and more, the machine will become more and more understand us. In Zhang Yixi's view, large data on the "God" of cyberspace, overlooking everyone.

"Although the IQ of machines is not necessarily higher than that of knowledgeable people, there is a limited amount of memory and limited data processing." The machine can not be tired, every moment can enter massive data. By perceiving and understanding the wisdom of the machine, it is possible to scale the gap of information. ”

Online less than three years, 220 million users, 20 million days live. For the team's achievements, Zhang Yixi is not too complacent, after all, today's headline algorithm pushed behind there are too many need to improve and optimize the place, the product is still being the user constantly spit slot. More importantly, January 19, Sogou has just launched a "micro-letter headline", competitors, are closely followed.

Finally, a data map of two, small partners please self-condemnation oh ~

(Responsible editor: Mengyishan)

Related Article

Contact Us

The content source of this page is from Internet, which doesn't represent Alibaba Cloud's opinion; products and services mentioned on that page don't have any relationship with Alibaba Cloud. If the content of the page makes you feel confusing, please write us an email, we will handle the problem within 5 days after receiving your email.

If you find any instances of plagiarism from the community, please send an email to: info-contact@alibabacloud.com and provide relevant evidence. A staff member will contact you within 5 working days.

A Free Trial That Lets You Build Big!

Start building with 50+ products and up to 12 months usage for Elastic Compute Service

  • Sales Support

    1 on 1 presale consultation

  • After-Sales Support

    24/7 Technical Support 6 Free Tickets per Quarter Faster Response

  • Alibaba Cloud offers highly flexible support services tailored to meet your exact needs.