Overfitting

  

Categories: Metrics

Our cousin Jaime has put on a little weight recently, and as we sit across from him at a family dinner, we can’t help but notice that his favorite shirt is stretching apart at the seams trying to accommodate his expanding girth. We snicker and tell him it looks like his body is overfitting his clothes. He snickers back and tells us we used the word “overfitting” incorrectly.

He’s right, we did. “Overfitting” is actually what happens when we create a model based on data that can’t be extrapolated to the real world. It can’t predict; it can only draw conclusions based on the historical data upon which we built it. And if that data—our training data—isn’t reflective of reality, then our model won’t be, either. This usually happens as a result of one of two things: either we have way too many data parameters, or we have so few data points that our model treats the anomalies as part of the pattern it’s supposed to recognize and predict.

As an example, let’s get back to poor Jaime. Let’s say we’re trying to predict how many cans of Pepsi he’ll drink in a given day, so we keep a log for seven days. On Monday, Tuesday, and Thursday, he drank four cans. On Wednesday and Saturday, it was five. Friday was three, and Sunday was zero. Based solely on this data, our model might predict that Jaime won’t drink Pepsi on Sundays. In reality, though, he was sick that day and only drank green tea with honey and milk. That anomaly should not be a part of our Pepsi predictions, because it was just that: an anomaly. Our model overfits the data.

Related or Semi-related Video

Finance: What are Financial Projections?96 Views

00:00

Finance allah shmoop what Our financial projections Well they're guesses

00:07

okay educated guesses we presume data is assembled reviewed and

00:12

then a crystal ball is gazed into And while projections

00:15

are made in the case of mature companies with long

00:19

cycle businesses the projections are often extremely educated and accurate

00:27

Yeah think about boeing's projected jet engine sales Well the

00:32

company takes three to four years just to set up

00:34

a fabrication run of a new style of engine and

00:38

they're working with rolls royce to do it Rolls royce

00:40

makes jet engines along with cars Their orders come in

00:43

and while they were partly paid for already and the

00:46

company knows it will produce somewhere between two hundred two

00:49

hundred fifty engines next year two hundred fifty two three

00:53

hundred engines following here and three hundred three hundred fifty

00:56

engines the year after that And yes there is some

00:58

variability and revenues But if the company produces fewer engines

01:02

well they'll have fewer costs as well So the range

01:05

they predict for profitability is pretty narrow That is they'll

01:09

have operating profit somewhere between eighty and ninety million bucks

01:14

each year Going up a little bit the next few

01:17

years That's The financial projection anyway can a bomb go

01:20

off in the factory or rather someone is smoking around

01:23

the fuel depot and then you know glam Yep that

01:26

can happen Is it likely No but it can And

01:29

it would really throw off the projections Right So projection

01:32

Is just a projection it's not a guarantee it's a

01:34

guess with other companies while like much younger ones projections

01:38

are well way more of a guess At the other

01:41

end of the rainbow from the boeing rolls royce jet

01:44

engine thing there's a brand new company just called rolls

01:47

it's an e bike company with assisted pedaling so that

01:51

riders feel like they're somewhere between superman and lance armstrong

01:56

You know the steroid guy company has had early success

01:59

having sold two thousand bike models off their website for

02:03

two grand each they're about to raise twenty million dollars

02:06

in venture capital funding Tio go big so now the

02:09

projections get really vague Will the company sell a million

02:12

bikes in three years Well that's what the founders think

02:15

but they're enthused and young for the company only sell

02:18

ten thousand bikes in three years Sure if they meet

02:21

the former projection they're billionaires If they meet the ladder

02:25

they're bankrupt Either way they can project with certainty that

02:29

their outcome begins with a b generally speaking the more

02:32

mature the industry and product the narrower or more accurate

02:36

the financial projection with startups who have no history Well

02:40

one never knows if they're projecting growth in the future 00:02:43.065 --> [endTime] like google or shmoop Come on

Up Next

Finance: What is the Dividend Discount Model?
2 Views

What is the Dividend Discount Model? Valuation of stocks can take numerous forms. One way in which one perspective on undervalued stocks is determi...

Finance: What is a WACC Model?
18 Views

WACC is an acronym for weighted average cost of capital. A company can raise money either through selling equity or by raising debt. When measuring...

Finance: What is the Black Scholes Model?
11788 Views

What is the Black Scholes Model? The Black Scholes Model is used to determine the price of call options. It looks at the change in stock price over...

Find other enlightening terms in Shmoop Finance Genius Bar(f)