Would you like to inspect the original subtitles? These are the user uploaded subtitles that are being translated:
1
00:00:00,002 --> 00:00:02,004
- [Instructor] If you analyze business data,
2
00:00:02,004 --> 00:00:05,007
and especially if you perform any kind of simulation,
3
00:00:05,007 --> 00:00:08,009
it's useful to know about the Poisson Distribution.
4
00:00:08,009 --> 00:00:09,008
In this movie,
5
00:00:09,008 --> 00:00:12,006
I will describe the characteristics of that distribution
6
00:00:12,006 --> 00:00:16,002
and also show you how to use it in Excel.
7
00:00:16,002 --> 00:00:20,003
The Poisson Distribution uses an average called Lambda
8
00:00:20,003 --> 00:00:22,008
and it determines the amount of time
9
00:00:22,008 --> 00:00:25,004
between events occurring.
10
00:00:25,004 --> 00:00:28,004
So for example, if you have customers
11
00:00:28,004 --> 00:00:30,007
that arrive on average every seven minutes,
12
00:00:30,007 --> 00:00:33,003
then this is what the curve would look like
13
00:00:33,003 --> 00:00:36,001
given that your data is distributed
14
00:00:36,001 --> 00:00:39,001
according to the Poisson Distribution.
15
00:00:39,001 --> 00:00:40,007
To show what that looks like in Excel,
16
00:00:40,007 --> 00:00:43,009
I will switch over to our workbook.
17
00:00:43,009 --> 00:00:45,005
I've switched over to Excel
18
00:00:45,005 --> 00:00:48,008
and my sample file is 04_04_Poisson
19
00:00:48,008 --> 00:00:52,001
and you can find that in the Chapter Four folder
20
00:00:52,001 --> 00:00:54,005
of the Exercise Files collection.
21
00:00:54,005 --> 00:00:57,003
And this distribution is called the Poisson Distribution
22
00:00:57,003 --> 00:01:01,006
because it was created, or at least discovered,
23
00:01:01,006 --> 00:01:05,003
by a French mathematician by the name of Poisson.
24
00:01:05,003 --> 00:01:08,002
The idea is that you want to calculate
25
00:01:08,002 --> 00:01:11,007
the probability of specific gaps between events,
26
00:01:11,007 --> 00:01:13,005
such as customers arriving.
27
00:01:13,005 --> 00:01:17,009
I've already set up a worksheet with a number of values
28
00:01:17,009 --> 00:01:20,004
and also my mean, or Lambda, which is seven.
29
00:01:20,004 --> 00:01:22,008
You can see that in cell E1.
30
00:01:22,008 --> 00:01:25,007
So, now I can calculate the probabilities
31
00:01:25,007 --> 00:01:29,005
of each of these gaps between events.
32
00:01:29,005 --> 00:01:35,007
So, I'll click in cell B2 and type an equal sign.
33
00:01:35,007 --> 00:01:36,008
And as you might guess,
34
00:01:36,008 --> 00:01:39,008
based on the way that functions are named,
35
00:01:39,008 --> 00:01:45,009
we'll use Poisson, P-O-I-S-S-O-N dot D-I-S-T.
36
00:01:45,009 --> 00:01:47,005
We need three pieces of information.
37
00:01:47,005 --> 00:01:52,008
The first is our X value that we're comparing. That's in A2.
38
00:01:52,008 --> 00:01:56,003
Then a comma. The mean is in cell E1.
39
00:01:56,003 --> 00:01:58,003
And I don't want that reference to change,
40
00:01:58,003 --> 00:02:02,000
so I will press F4, that's on Windows.
41
00:02:02,000 --> 00:02:03,009
If you're on a Mac it's command T.
42
00:02:03,009 --> 00:02:05,001
Then a comma,
43
00:02:05,001 --> 00:02:08,002
and we want the probability mass function
44
00:02:08,002 --> 00:02:10,001
or point probability.
45
00:02:10,001 --> 00:02:11,007
So we want to know the probability
46
00:02:11,007 --> 00:02:14,004
of an exact value occurring.
47
00:02:14,004 --> 00:02:20,003
So I will use the down arrow key to highlight false.
48
00:02:20,003 --> 00:02:24,007
Press tab, right parentheses, and enter.
49
00:02:24,007 --> 00:02:27,006
So we can see that the probability of a customer
50
00:02:27,006 --> 00:02:32,001
arriving at the same time as another customer is very low.
51
00:02:32,001 --> 00:02:37,000
So it's less than 1/100 of a percent.
52
00:02:37,000 --> 00:02:39,007
We can now create the same calculation
53
00:02:39,007 --> 00:02:41,008
for the other values in my list.
54
00:02:41,008 --> 00:02:43,009
So, I will click cell B2
55
00:02:43,009 --> 00:02:45,005
and then move my mouse pointer
56
00:02:45,005 --> 00:02:47,008
over the bottom right corner of the cell
57
00:02:47,008 --> 00:02:51,000
and double click the fill handle.
58
00:02:51,000 --> 00:02:52,007
And there we see the probability.
59
00:02:52,007 --> 00:02:55,004
So, we have our mean of seven,
60
00:02:55,004 --> 00:02:56,008
which occurs here at the top
61
00:02:56,008 --> 00:03:00,003
and will happen about 15% of the time.
62
00:03:00,003 --> 00:03:04,006
It's the same for six minutes between occurrences
63
00:03:04,006 --> 00:03:05,005
in this case.
64
00:03:05,005 --> 00:03:07,005
And you can see the other probabilities
65
00:03:07,005 --> 00:03:10,002
according to this curve.
66
00:03:10,002 --> 00:03:12,009
You can also determine the likelihood
67
00:03:12,009 --> 00:03:17,008
of a specific gap between customers or less occurring.
68
00:03:17,008 --> 00:03:19,004
And to do that, all you need to do
69
00:03:19,004 --> 00:03:23,001
is change the last argument from false,
70
00:03:23,001 --> 00:03:25,003
which will give you the point probability,
71
00:03:25,003 --> 00:03:29,006
to true, which gives you the cumulative probability.
72
00:03:29,006 --> 00:03:33,006
So if I make sure that B2 is selected,
73
00:03:33,006 --> 00:03:35,003
I can double click it,
74
00:03:35,003 --> 00:03:41,002
and then I will edit the last value
75
00:03:41,002 --> 00:03:44,007
or the last argument from false to true, enter.
76
00:03:44,007 --> 00:03:48,000
I get the same value because I'm only looking at zero.
77
00:03:48,000 --> 00:03:51,004
And then, I will click cell B2,
78
00:03:51,004 --> 00:03:54,001
double click the fill handle,
79
00:03:54,001 --> 00:03:57,000
and there we see the probability.
80
00:03:57,000 --> 00:04:00,004
So as you can see, it's rare that we will get values
81
00:04:00,004 --> 00:04:01,009
between zero and five.
82
00:04:01,009 --> 00:04:04,001
Those will happen about a quarter of the time.
83
00:04:04,001 --> 00:04:09,008
And then up to 10, because remember our average is seven.
84
00:04:09,008 --> 00:04:13,008
We have a probability cumulatively of 0.9
85
00:04:13,008 --> 00:04:16,000
and then as we get further out,
86
00:04:16,000 --> 00:04:19,008
the values get closer and closer to one.
87
00:04:19,008 --> 00:04:22,004
Two final things to note about the Poisson Distribution.
88
00:04:22,004 --> 00:04:25,009
The first is that it only deals with whole numbers,
89
00:04:25,009 --> 00:04:28,008
so you won't get any decimals in there.
90
00:04:28,008 --> 00:04:30,005
And secondly, you can think of it
91
00:04:30,005 --> 00:04:33,006
as the inverse of the exponential function.
92
00:04:33,006 --> 00:04:36,002
Remember when we used the exponential function
93
00:04:36,002 --> 00:04:39,006
we needed to divide one by Lambda.
94
00:04:39,006 --> 00:04:43,000
In the Poisson Distribution, we use Lambda by itself.
7333
Can't find what you're looking for?
Get subtitles in any language from opensubtitles.com, and translate them here.