UNIVERSIDAD AUTONOMA DE TLAXCALA

Facultad de Ciencias básicas Ingeniería y Tecnología

Probabilidad y Estadística
Dr. Eduardo López López

Semestre :2ro Grupo:” C”

Erandeni Yosselin Barrera Mateo

In this article, he tells us that the concept of randomness is not easily understood, as most people feel that any series
produced at random must contain very few, if any, extended runs of the same event and should represent the expected
frequencies of long-term events.

In the randomly generated series they can show none of these properties in the short term, when iPod users notice that
their playback function would alert seems to violate the properties when choosing, this shows an aversion to races in
studies where participants are asks them to generate random binary series, such as coin tosses these series that they
create tend to have too many flips and too few streaks of two or three hits than they would from a random process
(Bakan 1960; Orr, Federspieln and Maxwell 1972; Diener and Thompson 1985).

For an infinite number of coin tosses, the limiting ratio of heads equals the limiting proportion of tails. Many people also
expect this to be true in the short term, that is, relatively few flips of a coin. The belief is that short-term occurrences
other than this expectation must be corrected quickly, the belief that the occurrence of a random event will influence
your future events. An example is when a player has thrown ten heads in a row, he may believe that the tails are "due",
on the contrary, he may feel that the faces are "hot".

An example of the relative surprise of coincidences is the combination known as the birthday paradox. When asked,
"What is the probability that two or more people in a group of 25 will share the same birthday?" Most people find it
difficult to be that the answer is more than 50%. This is so surprising because we tend to focus on our own birthday,
which has a personal meaning therefore the question becomes "What is the probability that someone in a group of 25
will share my birthday?" In this case, the probability is much lower: less than 7%. In this section, we develop several
probability models for different of a mixture under the assumption that the function fle is truly random. Most of the
probability models can be found in standard textbooks such as Ross (2006) and Wackerly,Mendenhall and Scheaffer
(2007).

The shuffle function of any digital music player works by taking a collection of songs, called a playlist, and producing a
permutation sun. Each song will appear in the random playlist only once and each random play of a playlist is
independent of all others. This type of shuffling is in contrast to a physical shuffling of objects (such as a deck of cards).

Let N be the number of songs in a playlist. The number of songs in the random playlist that the user listens to before
selecting a different playlist or rearranging the same playlist is denoted as n. While n ≤ N, n is almost always strictly less
than N for longer playlists. Each song in an N-song playlist can be categorized according to one of several groupings,
including, Artist, album, date the song was added to iTunes, whether the song was purchased from the iTunes Store, etc.
particular grouping (Artist, Album, etc.), let G i denote the number of songs in the playlist that belong to the i- group and
denote the number of groups.

Assuming a random mix, the probability that any of the N! random changes will occur is 1 / N!. However, if the user only
listens to the first n songs in the random playlist, the probability that any of the N! / (N - n)! random changes will occur
equals (N - n)! / N!.Many comments from Levy and other iPod users claim non-randomness in the shuffle feature based
on the number of songs from a particular group, such as Artist or Al-boom, that occur during the first n songs in a shuffle
playlist. . Assuming a random mix, for a particular grouping
In 2004, Apple, Inc. introduced a new low-priced iPod called the Shuffle, built entirely around the Shuffle feature. The
first Shuf-fle contained approximately 120 songs. Since most users have well over 120 songs in their digital music
libraries, Apple uses a feature called AutoComplete to randomly select enough songs from a user's digital library to fill
Shuffle's capacity. In his Newsweek Article and Book, Levy tried autocomplete and reported "The first few times ..., I
found some clusters on the chosen songs. More than once the 'random' playlist included three tracks from the same
album since there are over 3,000 tunes in my library, this seemed to defy odd numbers.

The random variable described in Levy's statement is the maximum observation of a multivariate hypergeometric
distribution (max (Y 1, Y 2, ..., Y g)) and the event described is the capacity problem this maximum value will be three or
more. Published research on the multivariate hypergeometric distribution The contribution focuses on the values of (Y 1,
Y 2, ..., Y g) producing the highest probability value, the mode of the distribution.

A review of the literature did not produce any results for the probability distribution of the maximum observation
distribution. The reason could be related to the difficulty of calculating probabilities for the maximum distribution
observation. Looking at a simplified case, assume the number of songs in the playlist is N = 3,000 and each album in the
library contains G i = 12 songs for a total of g = 250 albums. In this case, the maximum number of songs in any album. It
can range from 1 to 12. To calculate the desired probability that the maximum number of songs on any album is three or
more, we calculate the complement event; the probability the maximum number of songs selected from any album is
one and two.

Clearly, if you listen to an entire random playlist, listen to each song in the playlist only once. However, this scenario
almost never happens, as users will only listen to the first n songs in a N-song playlist before rearranging the same
playlist or choosing another. Assuming random playback, the probability that a particular song occurs in the first n songs
of an N song playlist is n / N. Assuming that each random play is independent, the number of mixes S that occur until a
particular song is played. But in the first n songs of the mix it has a geometric distribution with probability n / N

Using various probability models, we can explain much of the anecdotal evidence for a non-random random mention
mentioned in Levy's Newsweek article and book. however, the apparent over-representation of a particular artist early
in the random playlist cannot be refuted based on the odds only. This perceived overrepresentation of certain groups
early in a random playlist has been reported not only by Levy, but also by. IPod users on several different websites.
Usually groupings are by artist or album, but some users have claimed random play favors the most recently added
songs to your iPods or songs chased through Apple's iTunes store by songs from your CD collection.

The act of clicking a song in the random playlist without listening to the entire song does not change the order of the
songs in the current random playlist. Lowered the null hypothesis of a random mixture, the number of songs AND i of a
particular group that appears in the first n songs of a N-song playlist has the hypergeometric distribution in equation)