2
$\begingroup$

Like above, I'd like to know, what exactly a skewed dataset is.

The explanation from stats.se.com sounds to me more like what I call an imbalanced dataset?

What is the distinction?

$\endgroup$
2
  • 2
    $\begingroup$ I would say in general skewed just means an asymmetrical distribution around the mean of the probability function: Skewness. So I would rather use it for individual variables. However, if you say dataset you mean more than one variable? $\endgroup$
    – El Burro
    Apr 12, 2017 at 15:53
  • $\begingroup$ It's better to call distributions skewed and datasets imbalanced to avoid confusion, as you've discovered. $\endgroup$
    – Emre
    Apr 12, 2017 at 16:47

2 Answers 2

1
$\begingroup$

In the context of the link, a skewed data set is referring to a dataset with a class imbalance problem. They are trying to build a classifier, but they have many more negative examples than positive examples. It's not a very precise term, but I've heard to used in this context a few times.

$\endgroup$
-1
$\begingroup$

Skewed classes are scenarios where the number of observations belonging to one class are significantly lower than those belonging to other class. These occur commonly in electricity pilferage, fraudulent bank transaction etc. There are many techniques which can help to reduce this non-uniformity like SMOTE etc.

$\endgroup$

Your Answer

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

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