#type1type2

[angusfng-blog]

Type 1 Type 2 Errors

A type 1 error is the rejection of a true null hypothesis (false positive) while a type 2 error is the failure of accepting a false null hypothesis (false negative). To think about type 1 and type 2 errors I like to use the example of a drug that treats a disease.

Let' s say the drug doesn't actually work and has no effect on the disease. This is the null hypothesis. A type 1 error would be rejecting that null hypothesis when we actually should have retained it. So in this case, by rejecting it,  we're saying that the drug actually has an effect. But in reality the drug actually doesn't. This is a type 1 error.

A type 2 error would be the failure to reject a false null hypothesis. For this example, type 2 would be accepting that the drug actually did work. However we know that it doesn't. This is a type 2 error.

The chances of committing these two types of errors are inversely proportional, decreasing type 1 error increases type 2 error rate and vice versa. Type 1 errors can be controlled. The value of alpha is the probability of a type 1 error occurring and this threshold is usually set at 0.05 (significance level). By doing so we are accepting there is a 5% probability of identifying an effect when actually there isn't one.

[nomiabalone]

Type 1 and Type 2 Errors

https://www.khanacademy.org/math/statistics-probability/significance-tests-one-sample/error-probabilities-and-power/v/type-1-errors 

https://magoosh.com/statistics/type-i-error-definition-and-examples/ 

https://www.statisticshowto.datasciencecentral.com/type-i-and-type-ii-errors-definition-examples/ 

Null hypothesis = no evidence to accept change  

Type 1 Error 

·         False positive 

·         We incorrectly think we have significant evidence to reject the null hypothesis so we think there is a change 

·         Incorrect rejection of a true null hypothesis 

Example: null hypothesis = Earth is not at the centre of the Earth. 

·         We know that's true but without seeing it for yourself (because that is difficult), one may conclude that other planets revolve around Earth and the hypothesis is thus false 

·         You can avoid this by reading more scientific journals 

Type 2 Error 

This is the error of accepting the null hypothesis even when it is wrong – false negative 

·         Failure of rejecting a wrong hypothesis 

·         We incorrectly think we don't have significant evidence, leading us to accept the null hypothesis even if its not true 

·         Failure to reject a false null hypothesis 

Example: null hypothesis = most people DO NOT believe in urban legends 

Mainly people don’t however you are surprised to find that some people, mainly retired residents with limited use of technology still believe in them. You therefore conclude that some people DO and so you conclude that your hypothesis is false. However you also fail to prove that most people don't. Therefore you incorrectly fail to reject the false null hypothesis that most people do believe in urban legends. 

[z5161477]

Type I and Type II errors

This week we learned about the 2 different types of errors when testing a hypothesis.

Testing a Hypothesis

An example of this is when you have a crush on someone and wanted to ask them out but you’re not sure if they like you back.

Errors shouldn’t happen when your hypothesis and reality match.

Hypothesis: They like you back.

Reality: They like you back.

or

Hypothesis: They don’t like you.

Reality: They don’t like you .

2 different types of errors will happen if the two don’t match.

Type I error: also known as false positive, is when you reject a true null hypothesis.

In this case, your hypothesis is “they don’t like you” but in reality, they like you back. You would miss an opportunity of finding a partner in life by not asking them out.

Type II error: also known as a false negative, is when you accept a false null hypothesis.

If your hypothesis was “they like you” and you gather the courage to ask them out, but in fact, they hated you, you would feel bad.

Security

These 2 types of errors are present in all aspects of life, including security. In many systems, it is crucial to be able to identify these errors and find a way to minimise it (or raise the ‘power of the test’).

Examples:

One of the most common example in security is Identification of threats:

Anti-Virus software should not be overly aggressive and create false alarms for every file it scans, but also it should not let viruses or malware to be able to bypass.

A few more examples of this include facial recognition, DDOS prevention, human verification (using CAPTCHA),...

Increase the power of a hypothesis test

Minimising one type of error over the other will lead to an increase of the remaining error (minimising type I errors will increase the chance of having type II errors and vice versa).

Some ways to mitigate this is to improve the process, use larger samples, minimise the consequences,...

Overall, it is impossible to avoid these errors. One must be able to identify consequences when making such errors and having a plan to deal with the consequences in the case of it happening.

[cellophanedreams]

Finally reunited with this wild child albeit short lived. Love you so much sweetheart, always my best girl. Can't wait for your next Sydney 😉 or till the next time in Singaland! #281215move #slutzcentral #type1type2 #ride #rile #mcnuggets #trampstamp #choker #windy #tgifinally #longweekendeveryweekend #weekendvibes #nomondayblues #sydney #australia #singapura #singapore (at O Bar and Dining)