Questions tagged [logarithm]
The logarithm of a number is the power to which the base must be raised to get the number.
690 questions
4
votes
3
answers
272
views
How to explain different regression results when dependent variable is in log or in level
I am running n event study like DID analysis where my dependent variable is expressed in logs. The variable is the aggregate amount of sales from a given firm to a specific type of partner over a year,...
2
votes
1
answer
227
views
When to use a log transformation in a regression? [closed]
I am completing a regression on the impact of drinking on income and am stuck on whether or not to log income for the dependent variable. I originally planned to use it for percentage interpretation, ...
- 21
0
votes
0
answers
21
views
Interpretation of the coefficients [duplicate]
What is the common way of interpreting coefficient when the independent variable is log(x+1) transformed and dependent variable is log(y) transformed? I have transformed my x variable in this way due ...
- 1
0
votes
0
answers
52
views
Correct use of "Log-transform + 1" for ecological data? [duplicate]
I am conducting an occupancy study using camera trap data with the camtrapR and unmarked packages in R.
I encountered convergence issues in my prey–predator co-occupancy model, so I calculated the ...
0
votes
0
answers
50
views
How to interpret the coefficient if the dependent variable is the growth rate calculated from the log of the original variable?
I am getting a bit of conflicted information while looking up online, so was hoping if someone can help me:
Suppose one runs an OLS where there are two dependent variables, calculated as follows:
a) $...
- 33
1
vote
1
answer
84
views
Should I log-transform individual timepoint measurements or the absolute change score when my outcome is volume (cm³)?
I have a longitudinal dataset where my outcome is white matter hyperintensity (WMH) volume measured in cubic centimeters (cm³) from brain MRI, collected at baseline and follow-up for each participant.
...
- 425
4
votes
1
answer
106
views
Applying quantile regression to naturally log-transform dependent variable and interpret results as symmetrical percentage change
I am using quantile regression and I was wondering whether it is appropriate to apply a natural log transformation of the dependent variable and then interpret the quantile regression results as ...
- 91
2
votes
1
answer
139
views
How to linearize a nonlinear y ~ x relationship (log-like increase after certain value)
I have a scatterplot of a variable $y$ against $x$ (attached below). The relationship is roughly flat until about $x = 60$, and then $y$ starts to increase exponentially.
I would like to transform $y$...
1
vote
0
answers
103
views
why can't I use the R squared to compare two linear regressions I if apply the log transform?
Yes the regressions have the same number of coefficients. The difference is just that one uses log(Y) instead of Y. The only ...
- 11
6
votes
2
answers
340
views
Spline and Log-transformation - are they mutually exclusive or not?
Today I am dealing with a conceptual question.
Suppose I have a logistic regression model, with a binary response (dependent) variable, and a continuous predictor (say, age).
I want to analyse the ...
- 838
3
votes
0
answers
62
views
Theoretical Applications of Combinants in Probability
I've recently come across the concept of combinants while reading about probability theory. The Wikipedia article on combinants provides a basic overview but doesn't go into much any detail about how ...
- 10.1k
8
votes
1
answer
268
views
Distribution of $Y=-\log(X)$ when $X\sim \operatorname{Beta}(\alpha,\beta)$
I have been looking at the distribution of $Y=-\log(X)$ when $X\sim \operatorname{Beta}(\alpha,\beta)$, so $Y$ is supported on $\mathbb R_{\ge 0}$. Might this be called a negative exp-Beta ...
- 45.5k
1
vote
0
answers
11
views
Interpreting an exponentiated coefficient from a regression of a log-transformed outcome when coefficient is < 1 [duplicate]
Plenty of stuff online (see here and here) about what to do when you log transform an outcome (exponentiate regression coefficients) and how to interpret the coefficients (1-coefficient*100 = ...
- 2,345
0
votes
0
answers
167
views
Overfitting in a dual natural logarithm model
I have the following function which models the evolution of gas intensity in a mass spectrometer:
$$y = a \ln(p(t + 32)) - b \ln(q(t+30))$$
which, as JJacquelin has pointed out below, simplifies to:
$$...
- 263
0
votes
0
answers
57
views
Building Time Parameters for Lognormal Model
I am trying to figure out how to build the parameters for time for my lognormal model (building a velocity profile)
the model has N lognormal curves, and each has a t and t0 parameter. While I ...
