Several core elements of machine recommendation,

First, originality, as long as it is original, more than 9% of machines will recommend it. By originality, I mean that you wrote it all by yourself, not pseudo original. Pseudo original may not be recommended, but if it is too fake, it will not be recommended, which will affect your originality index. Therefore, if we can be original, we should be original. If we can't be original, we need pseudo original, but it is better to modify the content by more than 6%. As long as it reaches more than 6%, it is generally recommended.

Second, the title should be consistent with the content, and don't be a headline party.

As long as the above two core elements are achieved, your article recommendation is affirmative, which is the case in today's headlines. Other new media platforms are not necessarily the same, but they are similar, and all new media platforms like original content.

Third, the keywords in the content should be compact.

The machine will automatically list some keywords in your article. In fact, these keywords may have nothing to do with the meaning of your content, but the machine is a machine, and it is impossible to read people's thoughts and realms. You can only choose some keywords according to the set algorithm. For example, if your article is written about braised eggplant, the keyword may be braised eggplant if it is selected manually. Cooking methods, etc., but the machine will not choose this way. He may choose eggplant, wok, seasoning, etc., which is far from your theme. Generally, he chooses dozens of keywords from your articles, and then matches the readers' interests and hobbies, and may recommend them to people who sell iron pots to see your braised eggplant. After understanding this principle, we can have a mental preparation when writing an article, that is, we should mention our core keywords more in the content, so that the keywords can be compact and concentrated and get more accurate recommendations.

Fourth, the content covers a wide range of people

Fifth, the more reposts, the more recommendations

Sixth, the higher the click, the higher the reading rate, the more recommendations

Click rate is also a way to measure the popularity of articles, but there is no click rate index in today's headline index. Therefore, the click rate is not necessarily the consideration range of machine recommendation, but many of them are complementary and interrelated. For example, if you forward more, the number of clicks will naturally be high, and the click rate will be high. Forwarding will also drive it. Many authors often work hard on the title to attract users to click. It is right to attract clicks, but the machine also has an algorithm that is the time for users to read articles and the completion rate. This is also an important factor. If the click-through rate is high and the reading rate is poor, then your article will be judged as the title party, and points may be deducted. In addition, it will affect your income, because the poor reading rate means that fans have not seen the advertisement. Since they have not seen the advertisement, the advertisement will naturally not be divided.

In fact, there are many algorithms for machine to judge the recommendation mechanism of articles, such as whether the content is interesting, easy to understand, illustrated and so on. However, as long as we can achieve the above six elements, the number of recommendations can definitely go up, and millions of articles are not a problem.

My name is Pang Zhang, and you are welcome to pay attention to my public service number. Top writers are fat, and there will be explosive software to send back.