#statistics #math #theory #centraltendency #estimative
- The first dimension of the data
-
Use a unique value to represent a collection/set of values
- This unique value can be used as baseline for other models
- Strong tendency or weak tendency is dependent on how dispersed are the data
-
Some measures of central tendency are:
- average/mean: Can suffer from outliers if they have extreme values
- median: It is robust, avoid outliers because use only the middle values
- trimmed mean: It is possible to remove the extremes, avoiding outliers as well, but using part of the population
- mode: Can be used if the frequency has a peak
- weighted mean: when you have different group and want to assign different weights for them. Ex: a sensor not working well can receive small weights
Insight
Given a measure of statistics Dispersion, find a central tendency which minimize the variation.
Q: Maybe can have something related to reduction of error on machine learning?
Reference
- Wikipedia
- Bruce, 2020, p4-8