Class Notes (1,100,000)
US (490,000)
UC-Irvine (20,000)
PSYCH (500)
Lecture 14

PSYCH 9A Lecture Notes - Lecture 14: Brett Hundley, Kobe Bryant, Euclidean Distance

Course Code
Bruce Berg

This preview shows half of the first page. to view the full 2 pages of the document.
Experimental Psychology
Fall 2017
Post-Midterm 2: Study guide for Lecture 18
1. Shepard scaling
2. Class demo: 2-distances (+ 1 binary judgment) between cities sufficient to place a city in a 2D
map; Demo: build intercity map, LA, Jackson, NYC, SFO
3. Monotonic easue of siilaity sufficiet do’t euie etic distaces to deie a representation of
items as points in a multi-dimensional Euclidean space
4. Good outcome: monotonic function between Euclidean distance in 2D and judged similarity
Data reduction example: 101 cities have 101x100/2 = 5050 intercity distances, can represent all these
with 2x101=202 x,y coordinates of the cities in a 2D x,y space. 202/5050 = .04, a 96% saving
5. Representation of similarity matrix in 3D space (pitch spiral)
2D space examples:
red color chips differing in lightness and saturation (not in hue)
color circle of chips differing primarily in hue (minimal differences in lightness or saturation)
Morse code
Various representations of speech sounds (syllables differing in initial consonant; 16 English
6. Measue of goodess of fit is “tess
(Omit Non-Euclidean Representations)
8. Additive trees (e.g. for animal similarities)
9. Nonmetric clustering
Shepard: consonants; body parts; Lecture: languages (number of common words in 100 most
feuet ods; Dai’s eolutioay tree from genes similarities
find more resources at
find more resources at
You're Reading a Preview

Unlock to view full version