2
$\begingroup$

Thank you for reading my question.

I used the clustering method to classify the age of the houses. This graph show the clustered results.

enter image description here

I should to get the age range, mean age, and median age for each age phase. I calculated them: age range=(maximum age of each phase)~(minimum age of each phase), mean age=sum of ages of each phase/the number of specimens in each phase, median age=find the value in the middle.

However, I unsure it is the right way to get them because there are outlier values. For example, in phase 4, one specimen had 39 yrs. Due to the outlier, the range of the phase 4 was 39-49 yrs.

I wonder whether I trimed the values to remove the outlier before calcaulting the age range, mean age, median age or used the orginal dataset including outlier mentioned above.

Thank you for sharing your advice.

$\endgroup$
1
  • $\begingroup$ What is the problem with just using the mean or median for each phase? $\endgroup$
    – Henry
    Jul 18 at 9:23

2 Answers 2

6
$\begingroup$

Ranges, minima and maxima are of course very sensitive to extreme values. (I prefer this term over "outliers", because "outlier" suggests that the data point is atypical - it may be completely valid, just very large or small.)

You can certainly calculate trimmed or winsorized means, or instead of the range calculate the 10% and 90% (or similar) quantiles. What is a "good" summary of your data depends on your field, and on any conventions your audience may use.

$\endgroup$
3
$\begingroup$

You may be able to summarize the data well enough for your purposes with the median and the interquartile range, which are both robust: insensitive to outliers.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.