Skip to main content

Questions tagged [compositional-data]

Refers to variables representing fractions of a total, i.e. all lying in $[0,1]$ interval and necessarily summing to one. Analysis of such data is often called compositional data analysis.

Filter by
Sorted by
Tagged with
0 votes
0 answers
9 views

Significance of factors/regressors on compositional data

Let's assume that I have composite data, similar to the "ArcticLake" example dataset in the DirichletReg package in R. Now, I would like to know whether various numeric variables and factors ...
Anti's user avatar
  • 159
1 vote
1 answer
55 views

Estimated marginal trends after regression of isometric log ratio transformed compositional data

I have longitudinal compositional data with four base response variables. I would like to compare the slopes of groups across time, but not for the ilr-transformed composition. I wish to compare the ...
Bryan's user avatar
  • 1,099
0 votes
0 answers
55 views

Do I understand correctly how to perform sequential binary partition?

I’m dealing with geochemical data. Let’s say I have the results of chemical analysis of a mineral with elements: A, B, C, X, Y, Z. X, Y and Z are elements belonging to the same group (let’s say they’...
Mike Surowski's user avatar
1 vote
1 answer
36 views

Inference from profile comparisons

What would be a good test to compare patterns of differences among treatments? For example, the Y maze has an outcome of % of entries into each of three arms. Thus, three numbers are generated for ...
Bryan's user avatar
  • 1,099
0 votes
0 answers
50 views

Multivariate analysis with lmer and lmerTest

I'm dealing with compositional data (data that sum to 1). They are inherently multivariate. One way to analyse compositional data is with a ilr (isometric log-ratio) transformation. I'm following the ...
Nee's user avatar
  • 11
0 votes
0 answers
33 views

Compositional data - interpreting Dirichlet regression with interactions

I did a Dirichlet Regression to assess the effect of independent variables on a compositional dependent variable. However, as I have 3 predictors, I do get interaction. I have a hard time ...
Nee's user avatar
  • 11
1 vote
1 answer
65 views

Analyzing compositional data (sum of proportions = 1) using mixed models with explanatory variables for each proportion

I aim to investigate how the relative abundance of species across communities is associated with the functional traits of each species. For each location ($>250$), I have compositional data that ...
Ruben's user avatar
  • 111
3 votes
1 answer
203 views

R - multinomial logistic regression with relative frequencies as response variable

My colleagues observed in an experiment involving categorical and continuous independent variables, how the species composition changes. Approximately equal numbers of microbes were used in the ...
Anti's user avatar
  • 159
3 votes
1 answer
41 views

How do I compare explanatory power of Dirichlet regression with other models?

I have a dataset with 7 variables expressed in % of the total (i.e. one of the constraints is that sum of the variables = 100%). I am looking at Dirichlet regression in R for the first time (I am ...
Ben Thuriaux-Aleman's user avatar
1 vote
0 answers
52 views

Principal Component Analysis (PCA) the constant sum constraint [closed]

I have data that sums to 100 % +/- 5%, and I was reading that data that sums to 100% should be altered before doing PCA. A method that I found others have done is to use the centered log-ratio (clr) ...
Stefg7's user avatar
  • 11
1 vote
0 answers
15 views

Comparing distribution of continuous data and binned data that are compositional

The Data I have a list of N ∼ 6700 population members who have a condition and are labeled with age in whole years and biological sex. For the parent population, I have The total number of each sex (∼...
abalter's user avatar
  • 1,128
0 votes
0 answers
40 views

"Reference" Continuous Variable / Variables that are perfectly multicolinear and sum to 1

I am conducting an ecological study that has data on municipalities as a sample. In this data, I have bins with proportions of individuals within a certain category that resides there, for example, % ...
Válmer Munhoz's user avatar
1 vote
0 answers
89 views

Scaling and centring in PCA of compositional data

I am following this review's approach for PCA using compositional data. It involves computing the centred log-ratio (CLR) transformation of the compositional data, and then running PCA on the ...
fm361's user avatar
  • 103
1 vote
0 answers
65 views

Compositional data analysis - counts and proportions

I have a dataset where I counted monkeys in 10 plots and assesses the proportion of 5 different habitat types covering the plot. The sum of the habitats per plot is always 100. I would like to test ...
CMa's user avatar
  • 11
0 votes
0 answers
31 views

Any reasons I shouldn't calculate the difference between CLR transformed variables to analyze my data in a time-independent way?

First time posting, so apologies for any missing info and the like. I have some microbiome data collected from two different treatment groups over two timepoints. I want to look at the compositional ...
Kam's user avatar
  • 1
0 votes
0 answers
352 views

How to handle zeros before carrying out CLR transformation to deal with compositionality before Spearman correlation?

My data is very zero inflated - even when carrying out previous filtering, such as getting rid of data that has very low counts and have values in more than 60% of samples. However, the 0 values still ...
aim6789's user avatar
  • 77
1 vote
0 answers
48 views

Can I transform the difference between two isometric log-ratio balances to interpret is as the ratio between two geometric means?

Suppose I have several compositions of 4 parts $A_i, B_i, C_i,$ and $D_i$ with $i = 1, \ldots , n$ taken from two locations ($X$ and $Y$), which are coordinates in the simplex $S^4$. I apply the ...
Tyler Sagendorf's user avatar
1 vote
0 answers
34 views

How to model the additive components of a random variable whose value is known

I would like to model the variables $Y_1, Y_2, …, Y_n$, which satisfy the constraint $Y_1 + Y_2 + … + Y_n = Y$; where $Y$ (or at least an accurate estimate $\hat{Y}$ thereof) is readily available. ...
Sebastian Chejniak's user avatar
0 votes
0 answers
28 views

Statistical analysis of compositional data - what test to use?

For one of my studies, I am playing with what seems to be a compositional data. In short, I have some signal that I decompose in 4 frequency bands (B1 to B4) in a way that I get the relative ...
Nee's user avatar
  • 11
0 votes
0 answers
25 views

How to use partial correlation on compositional (counts) data? (e.g., as described in Erb et al. 2020)

I've read through this paper: https://www.sciencedirect.com/science/article/pii/S2590197420300082 Which introduced me to the concept of partial correlation. I typically use Proportionality when ...
O.rka's user avatar
  • 1,472
2 votes
0 answers
34 views

Are there any compositional data models applicable to panel data?

I have a dataset spread across N departments and T years. I am interested in how voting shares change as a reaction to some regressor X. The dependent variable Y is vote shares: every department casts ...
a nna's user avatar
  • 21
1 vote
0 answers
28 views

Constant sum suvey data!

I have conducted a survey where respondents had to allocate 100 points over 3 choices. There were 8 questions like this in total, where the choices were the same but the attributes on the choices were ...
Druppelwater's user avatar
2 votes
0 answers
69 views

Am I correct that removing the intercept allow you to analyse compositional data?

I've heard that one solution to analyzing compositional data (in my case five predictors that are proportions summing to 1) is to simply remove the intercept. This seems like a much simpler solution ...
Dave's user avatar
  • 2,651
0 votes
0 answers
57 views

Regression with proportion values in the independent variables

I want to perform a regression where my independent variables all sum to 1. The independent variables are proportions of money invested in different categories. What I have done: When checking the ...
JoseMM's user avatar
  • 1
1 vote
1 answer
66 views

How to test differences (over time and between treatments) of a specific species in DNA metabarcoding sequencing data?

I have DNA metabarcoding sequencing data in the following format: plot Time_point reads_species_A Reads_species_B reads_species_C 1 T1 0 245 65 2 T1 48 455 0 3 T1 15 5 10 1 T3 153 23 564 2 T3 ...
Robvh's user avatar
  • 113
1 vote
0 answers
106 views

Dirichlet for panel data: correlated random effects?

I was wondering whether I can apply a Dirichlet regression model to panel data (using either dirifit in Stata or DirichReg in R)?...
Charlotte's user avatar
  • 123
0 votes
0 answers
569 views

Understanding why CLR transformed compositional data should not be used with correlation analysis

A colleague of mine asked me why use proportionality when they could instead use CLR transformation followed by Pearson/spearman/biweight midcorrelation. I vaguely remember reading why correlation ...
O.rka's user avatar
  • 1,472
6 votes
1 answer
501 views

Can I use logistic regression when all of the regressors sum to 1?

Let's say I want to perform a logistic regression (binomial) as : X ~ P1 + P2 + P3 + P4 + P5 where X is binary variable (0 or 1) and P1, P2, P3, P4, P5 are ...
Nicolas Rosewick's user avatar
2 votes
2 answers
379 views

Proper approach for modeling multi-class probabilities (proportions/compositional data)

I have a dataset in which the target values are counts of objects within each class, example (Xs are input features, Ys are values to be predicted): ...
Matek's user avatar
  • 991
0 votes
1 answer
36 views

Trying to find the appropriate statistical method(s) to answer a marketing question — analyzing composition of groups

I’ve got roughly 90k rows of data, where each one represents a customer at my company. One of the columns tells me the channel that brought that customer to my company. I’m looking at the channel mix ...
user366614's user avatar
1 vote
1 answer
59 views

Metric or Method to examine the correlation of a ratio

Let $X$, $Y$, $W$, and $Z$ be random variables where $X$ and $Y$ are different metrics of the same data point and $W$ and $Z$ are also different metrics of the same baseline datapoint. The assumption ...
TheNumber23's user avatar
0 votes
0 answers
39 views

Linking features via correlations

I am comparing two tables with different size of features (rows), but the same number of samples (columns). I know that a significant part of the features are the same in the two tables, but they are ...
Roger V.'s user avatar
  • 4,429
3 votes
1 answer
365 views

Compositional Data in R

I'm writing a work on the Aitchison geometry for compositional data and I have seen an Image I want to reproduce in R. I work with the "compositions" library and I want to understand how to ...
vitalmath's user avatar
  • 131
0 votes
0 answers
295 views

centered log ratio transformation of overlapping compositions

I have a total of $N$ units that are assigned to $D$ classes based on certain combinations of characteristics. If each unit is assigned to only one class, then the problem reduces to the usual ...
Alemu's user avatar
  • 125
2 votes
0 answers
570 views

How to fit a Dirichlet mixed effect model in R?

I have a compositional data from an arena experiment in which the duration of an animal in specific zones was denoted. The time in the arena was fixed for all individuals which makes the durations ...
Gretchen's user avatar
3 votes
1 answer
165 views

How does removal of symmetry (e.g. via constraints) in a Bayesian optimization search space affect search efficiency?

There are many examples of search space symmetry in real-world optimization problems in the physical sciences. To motivate this, here are some that come to mind: When optimizing a formulation such as ...
Sterling's user avatar
1 vote
2 answers
641 views

Compare paired before and after composition proportions

I have some data on the composition of purchases by households, before and after an intervention. This composition is known at the number of items or amount spent level within 7 categories, and can be ...
Stephen Clark's user avatar
1 vote
0 answers
156 views

What is the best imputation method for missing values in compositional data?

I'm a little newbie at statistics, so I'm sorry if my question is too dumb. I'm working with compositional datasets with several chemical elements (continuous data; arranged in columns), but also some ...
Concerned Student's user avatar
0 votes
0 answers
48 views

Compositional data: is it possible to make products of components/features in a composition?

I've never asked about Stack OverFlow, even though it's been so useful to me on many occasions! I'm sorry if this topic has already been discussed previously, perhaps I wasn't able to find the good ...
marideuzena's user avatar
2 votes
2 answers
770 views

Continuous independent variables that sum up to 100%

As I am studying the capital structure of firms, the total of proportions owned by different ownership types (free float, institutional ownership ... etc.) will be 100% of the company shares. In my ...
Mohammed Al-maghrabi's user avatar
0 votes
0 answers
161 views

Is it OK to transform non-normal proportional data

I have a dataset for 'time spent' (hours) carrying out 'behaviours' (walking, lying, standing, grazing), during day or night 'periods' of cattle. Time spent is my dependent variable, with behaviours ...
Kyle's user avatar
  • 21
0 votes
1 answer
285 views

Compositional data using R and difference between groups?

could you please help me in my research. I have a problem that can summarize: Is there a difference between stores in terms of the share of sales percentage in the 4th quarter. ...
aleksandar_m's user avatar
0 votes
0 answers
29 views

How to model choice among three quantities, subject to a constraint

I have a response variable that is the amount in monetary units (m.u.) that a person can choose to donate among three alternatives, subject to a constraint that the total amount should be equal to 100 ...
Manoel Galdino's user avatar
0 votes
0 answers
698 views

Log-ratio transformations, Aitchison Geometry, why logs?

I'm recently diving into the field of compositional data analysis and I'm still wondering why the log transformations that Aitchison proposes work. What I understand is that, having a $D$- part ...
James Arten's user avatar
0 votes
0 answers
277 views

Replacement of true zeros in compositional data

How does one go about replacing true zeros in compositional data sets? There is a lot of information on handling rounded zeros, such as multiplicative replacement (MultRepl in R), but not on true ...
Kyle's user avatar
  • 21
0 votes
0 answers
82 views

Compositional data analysis

I have one independent categorical variable "Period" (day or night) and 4 dependent continuous variables "Behaviours" (walking, lying, standing, grazing) which are compositional in ...
Kyle's user avatar
  • 21
2 votes
2 answers
439 views

GLM using proportions/percentage response variables

I have a dataset for time (hours decimal format) spent carrying out different behaviours of cattle in a day. I'm trying to run a general linear mixed model with two independent variables (lactation ...
Kyle's user avatar
  • 21
0 votes
0 answers
38 views

I'm looking for a statistical test more robust than the sign test for my experimental analysis

I made an exploratory experiment that consists of a pair of readings. I used Scanning Electron Microscopy-Energy Dispersive Spectroscopy (SEM-EDS) to read elements on several mineral surfaces, getting ...
Jean Braga's user avatar
2 votes
1 answer
78 views

Should I use Linear Regression in this case?

I'm doing a personal project on regression and I hope I can get some advice from you on several problems. The dataset I'm having is about cuisine, shape is 180x9, and the 2 continuous variables are &...
statslearner's user avatar
1 vote
0 answers
65 views

Using compositional data analysis to represent chemical compounds

I've recently got some insights about compositional data analysis, wondering whether it could be suitable for the framework I'm currently in. Recently, I've been very interested in trying to find some ...
James Arten's user avatar