How can you implement a random drop percentage?

For example, there are 4 numbers (maybe more), 10, 15, 40, 12, but 10 percent has a dropout of 20%, 15 = 25%, 40 = 5%, and 12 = 50%, maybe there is some then the library?

  • 3
    Yes, MatLab is easier to implement. Read about the types of distributions and how to generate them. But in the general case when accuracy is not so important (when you are not a nuclear physicist), you just have to write the function yourself. Occurs value from 0 to 100. if it is <20 => then it is 10, if <45 (20 + 25) then 15, etc. - Makarenko_I_V

3 answers 3

You need to fill the array with numbers so that the number of occurrences of each corresponds to the probability of its appearance in the output.

$a = array(10,15,40,12); $ver = array(20,25,5,50); unset($b); foreach($a as $key => $val) { for($i = 0; $i < $ver[$key]; $i++) { $b[] = $val; } } shuffle($b); printf($b[0]);

    We paint the coefficients (so that there are always percentages - we divide each by the sum of all and think what to do with accuracy) :

     $chancesMap = [ 10 => 20, 15 => 25, 40 => 5, 12 => 50, ];

    We write function:

     /** * @param array $chancesMap Value => Chance * @return mixed * @throws Exception */ function getRandomValue($chancesMap) { $chancesSum = array_sum($chancesMap); $randomValue = rand(1, $chancesSum); foreach ($chancesMap as $value => $chance) { $randomValue -= $chance; if ($randomValue <= 0) { return $value; } } throw new Exception('Не может быть'); }

    Checking:

     $testIterations = 10000; $results = array_flip(array_keys($chancesMap)); // запустим несколько раз.. for ($i = 0; $i < $testIterations; $i++) { $value = getRandomValue($chancesMap); $results[$value]++; } // проценты.. array_walk($results, function(&$count, $value) use ($testIterations) { $count = round($count / $testIterations * 100, 2); }); var_dump($results);

    It turns out:

     array(4) { [10]=> float(19.7) [15]=> float(25.13) [40]=> float(5.39) [12]=> float(49.84) }

      I do not know about the library, it is hardly there. This is done simply

       $a = array (10, 15, 40, 12); $prob = array (20, 45 /*20+25*/, 50 /*45+5*/, 100/*50+50*/) $r = rand(0, 100) /* включительно, желательно равномерное распределение */ for ($i = 0; $i<4; $i++) { if ($r <= $prob[$i]) return $a[$i]; }

      Sorry for the pseudocode, php I know very little.