Posts

Showing posts with the label Statistics

You shall know a word by the company it keeps.

Quote of the Week 2 " You shall know a word by the company it keeps." ―  John Rupert Firth Cited in countless publications on Natural Language Processing (NLP) including  Christopher Manning's Stanford 2022 " Lecture 1: Introduction and Word Vectors " (slide #18).

Precision is vanity, recall is sanity.

Quote of the Week 42 "Pr ecision is vanity, recall is sanity. " ―  Martin White Title of a recent blog article by Martin White .   For a lot of use cases in information management and search ("finding all the relevant results"), false positives (as a result of lesser precision) are easier to handle than false negatives (as a result of insufficient recall) as one might not even know these misses exist in the information base (" You don't know what you don't know "). For more information on Type I ("false positive") vs. Type II ("false negative") Errors see for example  Wikipedia .

Correlation does not imply causation.

Image
Quote of the Week 23 " Correlation does not imply causation. " ―  Any savvy statistician Triggered by too many news headlines and presentations that cite misinterpreted research studies' results & claims.  See for example this blog or Wikipedia for explanation. Source:   Lucy D'Agostino McGowan, " Hill for the data scientist: an xkcd story "

All models are wrong, but some models are useful.

Quote of the Week " All models are wrong, but some models are useful. " ― George Box