Why can't we just add numbers to compare incomes?
Image: NASA/GSFC/LaRC/JPL, MISR Team ,przeniósł na Commons Dobrzejest, wikipedia.pl : Adi4000[1], Public domain, via Wikimedia Commons
to use log-transform: when data is right-skewed or spans multiple orders of magnitude
Why can't we just add numbers to compare incomes?
Imagine you're comparing incomes from different countries, like the US, India, and Switzerland. The numbers are huge and don't match up easily.
By using the log transform, we can make these huge numbers smaller and more comparable. It's like leveling the playing field for a fair comparison.
Example
An American earning $50,000, an Indian earning ₹1,000,000, and a Swiss earning CHF 100,000. After log transformation, these incomes become more manageable and comparable.
Remember this
The log transform helps us compare incomes from different countries on a more equal footing.
Text adapted from Wikipedia, licensed under CC BY-SA 4.0.
Related concepts
log-probabilities are used instead of probabilities: avoids numerical underflow
Why can't we just add up tiny chances over time?
O(n log n) is the lower bound for comparison-based sorting
Ever wonder why sorting can't be faster than a certain point?
log-loss / cross-entropy loss penalizes: confident wrong predictions more heavily
How wrong predictions hurt more than they're supposed to
Likelihood function
Why can't we just guess probabilities correctly all the time?
Adam has bias correction: divides by (1-β^t) in early steps
Why do we sometimes need to fix mistakes in computer decisions?
Global Forest Change dataset
Global Forest Change dataset covers 2000-2024
Swipe through more Machine Learning concepts
Open Pocket Polymath