Algorithm, a judge for a competition's score when how to let the score to the real value can't deviate too big can't deviate too small? Deviation is too big or small deviation scores are problem, how to implement this algorithm, because regardless of order when the score

CodePudding user response:

Now diving not remove a remove a minimum points then average

CodePudding user response:

Isn't this removes a remove a minimum points judges all score compares in scoring with has been put in storage if the deviation is too big or small deviation of the evaluation of very directly to score will run to the administrator administrator will do processing

CodePudding user response:

The punishment mechanism is introduced to judge the score, when the judges score, according to the judges scoring peace divide variance, to automatic grading of the judges, and introducing the punishment mechanism, in this way, in order to avoid their score is too high, the referee will try to put their own scores to average,

CodePudding user response:

What is a "true value"?

I think the bentley this car is very ugly. Although price is high. But I don't feel good-looking appearance.. So I played six points

Others played 9.5. You can beat my score?

I think this is not reasonable.

And hypothesis according to the score to calculate so-called "deviation", then the first man to play 1 minute, the second man I 2 points and so on.. To 10 points..

I don't think this should be come out..

1 or 2 you calculate a numerical definitely low.. So the third person to play 10 is 8 to 9 points behind... Before is 2 people is to make trouble..

You are judged by has score, so there are two grades, you are invalid at the end that is low so that also is not reasonable...

So I think the algorithm itself isn't what's the problem?

CodePudding user response:

Least squares method, and then take a range, not in this range

CodePudding user response:

If you want to use the more advanced algorithm, consider the ant colony algorithm and genetic algorithm

CodePudding user response:

Tell from the rationality of
Score itself is subjective, and can't say because someone's score doesn't "fit in", then to judge his score, you said the differences in the scoring system with us is too big, please score again, what is the point that this score,
Tell from the implementation (I don't care, I would do such a function)
First you need to create a reference, as to how the reference comes from your system as a whole logic and business considerations, you can consult the scoring rules of sports (average) to high, low, also can produce, such as the score for history score then directly take average, when the score and average over a certain range (also need you define this divided into) tag when abnormal,

CodePudding user response:

Unrealistic ideas,

Statistics can only help you to complete, average, variance, the median, bell-shaped curve characteristics of this kind of distribution

And multivariate binomial regression, support vector machine (SVM), the envelope analysis, can only say that the packet data classification, picked out a good or bad, or ignore individual jump bigger

At most so, as to the true not true, don't know, as the upstairs said, if 9 adult deliberately not real, how can you?

As now the movie reviews and box office, true? Box office can play false, comments can infuse water, so you want to according to the data for real not realistic, because the people involved, someone involved in things, algorithm is only on the game theory is true, this is people play together their own self-interest, self decision data and results, but the result couldn't push the heart

CodePudding user response:

Like a company 9 adults in fish in troubled waters, then individuals being honest to do things, then you say that someone else's index obviously, unlike most people, you said "administrator" talk with them?

So, this is the game, see more is a good man, or a bad person, is a bad money drives good money, or good catch bad money, in the end is the heart of the game, instead of Numbers can play out

CodePudding user response:

Now called artificial intelligence, is to look for relative fair 100 w group, then put a characteristic according to the statistical regression, using the binary clustering characteristics do

It is now called "artificial intelligence"

CodePudding user response:

Determine deviates from the value of the difference between the first, and then sorting, finally, a sampler effective collection of data,