Latin Squares Kaelen Medeiros Content Quality Analyst DataCamp - - PowerPoint PPT Presentation

▶

latin squares

Latin Squares Kaelen Medeiros Content Quality Analyst DataCamp - - PowerPoint PPT Presentation

Jul 31, 2023 256 likes •461 views

DataCamp Experimental Design in R EXPERIMENTAL DESIGN IN R Latin Squares Kaelen Medeiros Content Quality Analyst DataCamp Experimental Design in R Latin squares Two blocking factors (instead of one) All factors must have the same number of

slide-1

SLIDE 1

DataCamp Experimental Design in R

Latin Squares

EXPERIMENTAL DESIGN IN R

Kaelen Medeiros

Content Quality Analyst

slide-2

SLIDE 2

DataCamp Experimental Design in R

Latin squares

Two blocking factors (instead of one) All factors must have the same number of levels Key assumption: the treatment and two blocking factors do not interact Analyze like a RCBD

slide-3

SLIDE 3

DataCamp Experimental Design in R

Latin square diagram

slide-4

SLIDE 4

DataCamp Experimental Design in R

Why is it a Latin square?

slide-5

SLIDE 5

DataCamp Experimental Design in R

Intro to NYC scores

nyc_scores is an NYC open dataset

Downloaded from Kaggle Includes: All accredited NYC high schools SAT scores (Reading, Writing, and Math) 2014-2015 school year

slide-6

SLIDE 6

DataCamp Experimental Design in R

Let's practice!

EXPERIMENTAL DESIGN IN R

slide-7

SLIDE 7

DataCamp Experimental Design in R

Graeco-Latin Squares

EXPERIMENTAL DESIGN IN R

Kaelen Medeiros

Product Data Scientist at DataCamp

slide-8

SLIDE 8

DataCamp Experimental Design in R

Graeco-Latin Squares

Three blocking factors All factors must have the same number of levels Key assumption: the treatment and three blocking factors do not interact Analyze like a RCDB

slide-9

SLIDE 9

DataCamp Experimental Design in R

Graeco-Latin Squares

slide-10

SLIDE 10

DataCamp Experimental Design in R

GLS - explanation

slide-11

SLIDE 11

DataCamp Experimental Design in R

Let's practice!

EXPERIMENTAL DESIGN IN R

slide-12

SLIDE 12

DataCamp Experimental Design in R

Factorial Experiments

EXPERIMENTAL DESIGN IN R

Kaelen Medeiros

Product Data Scientist at DataCamp

slide-13

SLIDE 13

DataCamp Experimental Design in R

Factorial Designs

2 or more factor variables are combined and crossed All of the possible interactions between levels of factors are considered as effects

n the outcome

Example: high/low water and high/low sunlight's effect on plant growth.

slide-14

SLIDE 14

DataCamp Experimental Design in R

Factorial Example

slide-15

SLIDE 15

DataCamp Experimental Design in R

2^k factorial experiments

2^k factorial experiments involve k factor variables with 2 levels It results in 2^k number of combinations of effects to test Analysed with a linear model and ANOVA Also use TukeyHSD() to determine which combinations are significantly different

slide-16

SLIDE 16

DataCamp Experimental Design in R

Let's practice!

EXPERIMENTAL DESIGN IN R

slide-17

SLIDE 17

DataCamp Experimental Design in R

What's Next For Experimental Design

EXPERIMENTAL DESIGN IN R

Kaelen Medeiros

Product Data Scientist at DataCamp

slide-18

SLIDE 18

DataCamp Experimental Design in R

What's Next?

Other factorial designs (besides 2^k) including fractional factorial designs Experiments with random factors Nested designs Split plot designs Lattice designs

slide-19

SLIDE 19

DataCamp Experimental Design in R

Go forth & design experiments!

EXPERIMENTAL DESIGN IN R

Core Hubs:

All Converters PDF Hub Video Hub Audio Hub Image Hub File Compressor