Random sampling
Referring to a simple Lottery "1 of 5". Suppose the numbers 1, 2, 3, 4, 5, dropped to 20 circulations with frequencies of 2, 2, 4, 8, 4. To create a table that presents the data in a usable form.
_ N _ _ F _ _ RF _ _ _ CRF
1 2 0.10 0.10
2 2 0.10 0.20
3 0.20 0.40 4
0.80 0.40 8 4
4 5 1.00 0.20
How can you choose the numbers to select the number of probability is directly proportional to its frequency of occurrence in history (or, equivalent, is equal to its relative frequency)? A very simple way would be to take your tiles set of 10 ornaments China, with pen on one of them to call a number one, the other two, two, three, four, and two 5:0 pm 4. Then put them in a jar, shake well, and blindly remove nodules. Probability, which would you choose 4 number dominoes is 0.40, the probability that you select a number of dominoes is 0.10, and so on. All nodes of the fingers of choice (the next election the junta extracted must be returned to the Bank, and that the unrest) is random with probability of the frequency selected on the number fell in previous editions. This-the random selection, but weighted random (probability-weight), with "hot" items weighing more than the weight of "cold" numbers. Otherwise you could do the same random sample, using uniform random numbers, and the following distribution function. Our distribution function:
_N_ _CRF_
1 0.10
2 0.20
3 0.40
4 0.80
5 1.00
Use RAND () to select one of two-digit uniform random number. Assume that this is-the number 0.34. Look at the column of CRF. The number fell between 0.34 and 0.40 0.20, choose the number 3 (You randomly selects 3 in distributors). Use RAND () to select another random number. Suppose that this time you have 0.69. From 0.69-between 0.40 and 0.80, select the number 4. And so on. What is happening here? It is very simple. Since the RAND () produces a uniform random number, then the probability that the number of falls-equal to 0.20 -0.40 0.20 (less 0.20 0.40), then there is the relative frequency of fall-3 emission (see the table at the top of this page.) Thus, the choice to interpret 0.34, a number between 0.20 and 0.40, which is similar to a random selection of numbers 3. Likewise, our random number is between 0.40 and 0.80 0.69, is interpreted as a randomly selected number 4:. The relative frequency corresponding to 4, 0.40, is the same as the probability that we could choose a uniform random number between 0.40 and 0.80 This random selection process can be automated (no conscience), using the following simple rules. Rule. Use RAND () to select a random number ravnomernoraspredelennoe, call it r. If R falls between two values of a column, the FRC, choice of column N is a number that corresponds to the higher of the two values of which obtained r. If R is not equal to one of the values in the column, column choice no CRF corresponding number. Illustration with the rule. If R = 0.53, picked 4 because 0.53-lies between 0.40 and 0.80, and 4 correspond to 0.80. If R = 0.31, choose 3 because it is between 0.20 0.31 and 0.40, and 3 corresponds to 0.40. If R = 0.20, choose 2. If R = 0.80, select 4. Lottery conducted by the Government as Lotto Texas use mechanisms of mixing and playing balls, trying to ensure that they totally random occurrence. By "completely random" I mean simply that each ball in a game full of balls, can be selected with equal probability in any quantity. This is a sample function of a uniform distribution. If, in fact, the machine reaches the perfect objective randomization, the RAND function (50) to the Texas Lotto or RAND (40) for the Plus Lotto-how a good a way as any to select the numbers for your tickets.