I thought about studying hard Xi I had nothing to do today, anyway, there was nothing to do, and I could learn as much as I could, so I took notes when I was studying Xi.
I hope my notes are helpful to everyone, and the content of the notes is relatively rough, because I just rushed to understand a little bit, so I didn't write in great detail.
Beginners can learn Xi with me!
The following notes are written:
Learn about the referral mechanism.
1. How is a piece of content recommended to users after it is published?
A: Directly match the current user with the content that interests them most.
2. How to judge whether users like it or not?
Answer: Learn from user actions, such as sharing, liking first, the length of time to read comments, etc.
3. What is the main basis for matching content to users?A: According to three factors: the user, the content, and the degree to which the user is interested in the content
4. How does the system understand the content we create?A: Extract keywords from content, or use AI technology to identify audio and specific content to quickly classify content.
5. So how do we provide users with the content they like?
A: The first is through portrait, age, gender, historical browsing content, environmental characteristics, etc
The second is to determine what the user is interested in through the user's actions
6. So what kind of life course will the content establish in the system after it is published?
Answer: There will be two links, one is the preliminary examination, and the other is the reexamination.
Only by passing the initial prison can we know whether this content will become popular, so there will be a cold start during the preliminary trial.
After the cold start, whether this content can be launched in the future depends on the strength of creation
In the process of recommendation, get customer evaluations, such as: users will click on him, comment on him, or even report. Based on this behavior, the content will enter the review process, which is the review.
In the review, if it is found that there are problems such as title party, cover party, vulgarity, falsehood, etc., the system will stop recommending.
The above are the notes learned today, do you understand?
We can make good use of Toutiaojun's recommendation mechanism to write what we want to write, if there is no traffic to write it, we will write it in another way, as long as we insist on writing, we will definitely succeed.
Too many examples tell us that perseverance is victory. Sometimes income is not important, only often, and in the end it will come naturally.
Let's work together!