[reading this article is enough!!! Easy to understand] confidence level understanding (95% confidence level and confidence interval)

Because I saw this concept when I read the paper , In many experiments, it can also be regarded as an evaluation index , But I don't quite understand . This is a concept in statistics , Although I have studied Statistics , But I don't remember the concept , Come back ,Ｏ(≧ mouth ≦)Ｏ
The technical terms of Baidu Encyclopedia are difficult to understand , I have integrated many people's explanations and my own understanding, hoping to describe this concept clearly in the most accessible words .
Although the space is a little long , But after reading it, I will really understand .
First , In statistics , What we all know is , The overall level can be assessed by sampling , It can also be said that the measured value is used to estimate the real value of the population . For example , I've always wondered how much time the citizens of our city spend on their mobile phones every day , Of course, it is impossible to ask the people of the whole city again , After all, 362.09 ten thousand people . Then we can take samples , Select some people to do the survey . Suppose it is within my ability , I can only investigate 100 personal , By asking, etc , this 100 Individuals play mobile phones on average every day 8h, Then I can say that through the sample survey, the average citizen of our city plays mobile phones every day 8 Hours ？ Of course not. , So if I take another random sample 100 Personally, the average time they spend playing mobile phones every day is 7h perhaps 5h Well ？ May I draw a direct conclusion ？ Obviously, the data is unreliable , There is a feeling of generalizing , Need to know , If it's true 362.09 Million people to investigate , There must be a real expectation , But exactly how much , I can't say , I don't know. . If there is another indicator , Or it can be understood as a constraint , ad locum , I define this indicator as ： Confidence in my findings , Set to several levels , That is, a little believe （30%）【 notes ： test 100 Time , Yes 30 Times contain real expectations , Not here 30 The value of this survey result is equal to the real expected value , Because there will always be errors , It is impossible to completely equal the final expectation , Just think that this expectation can't be known at all . And this 30 Next is 30 A confidence interval , The back can speak , So it contains expectations , contain ！】, Reluctantly believe （50%）, Believe in （70%）, I believe very much in （95%）. then , Make a difference between your test results and the real results , Compare this difference with the set probability , The formula is if the difference is less than the set probability , Suppose it is the above 95%, Then we can say , I have a 95% Grasp （ That is, I believe very much ） I think the result of my investigation （ The assumption is 8h） It is very close to the real expectation .
therefore ,95% Is confidence . So how does the confidence come from ？ Borrow a picture from someone else , I can easily understand it by applying my example to this diagram .

First understand the confidence interval ：
You can refer to this link , The confidence interval is reasoned in detail
https://www.zhihu.com/question/26419030
Then apply it again ：
The big dotted line is the real expectation 8h, The short line above is the confidence interval , Suppose I investigate 100 Time , this 100 In the second place 95 Time （ namely 100 There are two confidence intervals 95 A confidence interval ） Both include real values 8h, So confidence is 95%.

Then a netizen's answer can also help understand ： The confidence interval is a random interval . Random , It means that the endpoint is a random variable , This random variable is usually a statistic , When different samples are taken, they correspond to different values , Thus corresponding to different intervals . For some samples , The corresponding interval contains the true value of the parameter , Others do not contain . If in 100 Constructed in sub random sampling 100 Interval if 95 The second contains the parameter truth value , So the confidence is 95%.
link ：https://www.zhihu.com/question/26419030/answer/81409702
After reading the above, let's take a look at the concept ：
** Degree of confidence ：** Centered on the measured value , Within a certain range , The probability that the truth value will appear in this range . General settings 95%, Is the general confidence （ Confidence level ） Set value of .
confidence interval ： Under a certain confidence , Centered on the measured value , The range of true values . The value range of truth value under certain probability （ Reliable range ） It's called the confidence interval . Its probability is called confidence probability or confidence （ Confidence level ）.

After reading these , Must be able to understand , Let's look at the concept of technical terms , It's going to open up ！
Reference resources ：
https://www.zhihu.com/question/20183513/answer/15040378
https://wenku.baidu.com/view/cf67d1da360cba1aa811da23.html