Compositional Data Flashcards

1
Q

Compositional Data

A
  • non-negative vectors that sum to a unit-sum constraint
  • can be proportions, percentages, counts/frequencies
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Relative Property

A
  • only ratios/relative proportions between components are meaningful
  • total sum value is arbitrary - absolute scale does not matter
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Spurious Correlation

A
  • misleading, neagtive correlation between components due to sum constraint
  • change in one value of one composition, forces change in others
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Scale Invariance

A
  • ratios between components unchanged under rescaling

multiplying by constant - ratios stay the same

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Subcompositional Coherence

A
  • relantionships between parts remain valid even when analysing a subset of components

A,B,C - analysing just A,B shoudl not give conflicting conclusions

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Subcompositional Dominance

A
  • if one component dominates the full composition, it should dominate in any subcomposition

if A always greater than B in A,B,C - then should remain true in A,B

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Permutation Invariance

A
  • order of the components should not affect the analysis

A,B,C should give the same results as B,A,C

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Ternary Diagram

A
  • graphical visualtion of three components
  • near vertex - high concentration of that component
  • near centre - equal proportions of all components
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

ALR

A

additive log-ratio
* divides ratios by one component
* dependent on choice of divisor

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

CLR

A

centered log-ratio
* divides ratios by geometric mean
* covariance singular as all components retained - robustness issues as not fully independent

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

ILR

A

isometric log-ratio
* uses orthonormal coordinates to transform
* creates independent, orthogonal coordinates
* harder to interpret

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Rounded Zeros

A
  • represent values that fall below some detection limit
  • not true zero values
  • due to measurement error or below detection limit
  • treated by replacing the zero values
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Structural Zeros

A
  • true zeros
  • actual zero or absence of component
  • informative
  • carry important information
  • model-based approaches
How well did you know this?
1
Not at all
2
3
4
5
Perfectly