Questions tagged [fitting]
This tag is overused, so consider applying a more specific tag.
839 questions
2 votes
0 answers
85 views
F-test: Is the $\chi^2$ difference of nested models $\chi^2$-distributed?
I have recently started working with F-testing for determining optimal polynomial order of a fit. I have calculated the $\chi_k^2$ for a certain model with $k$ parameters and $\chi^2_{k'}$ with $k'>...
- 21
2 votes
2 answers
107 views
Loglikelihood of fit over power-transformed data (power $p$) constant accross $p$?
Just consider data x and its powers x^p. I wanted to check the impact of p on the goodness ...
- 435
0 votes
0 answers
82 views
Why does my gini index increase when my proportion of training data increases
I've been fitting a binary logistic extreme gradient boosted model using different random samples of the data as training and computing the Gini index (coefficient). However, as I increase the ...
- 808
0 votes
1 answer
68 views
DCC-GARCH for series with unit root
I have a question connected with DCC-GARCH models. Let's have K timeseries, some of them are I(1) series, some - trend-stationary (T) series. For TS time series I can substract trend directly (this ...
- 260
1 vote
1 answer
81 views
Fitting two datasets with different behaviors generate error for one of them
I have an issue for fitting multiple datasets using the same equation. I measure some parameters in two conditions: ...
- 123
0 votes
0 answers
58 views
Convexity of loss function in model fitting without known data
I am a bit confused about the concept of convexity analysis when doing model fitting. Say I have developed some model of two parameters $f(x;\theta_1,\theta_2)$, that I will plan to fit to some data I ...
2 votes
1 answer
92 views
Choosing the number of percentiles to be determined from dataset, for distribution fitting
If I have a dataset of $N$ values, and I want to fit a probability distribution to it, I would calculate the percentiles of the scores in the dataset. However, I am troubled by the number of ...
- 209
0 votes
0 answers
41 views
Comparing event frequencies in 4 treatments over time
I have an experiment with 4 treatments where I count the frequency of the same event across all treatments once per week and I have 26 weeks of data. The y-axis is % of subjects experiencing the event ...
2 votes
2 answers
233 views
Estimate of error for Google-Rating
Overview. For a private project of mine I am looking into Google's rating system. There you can rate e.g. a local store with 1, 2, 3, 4 or 5 stars. I obtained data from a rather large facility in my ...
- 145
2 votes
1 answer
88 views
What popular parametric distribution (or mixture of distributions) does a good job at modeling the age of death of humans?
For those who don't know, the distribution for the age that someone dies looks something like this: Obviously there's a spike close to age 0 that accounts for infant mortality. That can be easily ...
- 561
0 votes
0 answers
43 views
Why does GEV fit sometimes not fit the tails well?
I am performing a generalized extreme value analysis using about 20 years of data sampled every 1 minute. I am doing this in order to predict return levels at e.g. 1-in-50 and 1-in-100 intervals. The ...
- 947
1 vote
0 answers
73 views
How to get 1-in-100 return period from GEV when using a block length shorter than a year?
I am computing the Generalized Extreme Value distribution for a dataset containing about 15 years of data sampled every 5 seconds. I want to estimate the 1-in-50 or 1-in-100 year return level from the ...
- 947
3 votes
2 answers
401 views
Can't fit Gaussian Mixture Model, estimates wrong parameters
The test below generates samples from Gaussian Mixture Model, and then fits it back. The fit model is totally different from the original. Why? How is it even possible, the results are not just ...
- 583
2 votes
1 answer
85 views
R what is the best glmmTMB model family to fit positively skewed index data
I am attempting to create a linear mixed effects model (lmer) with an positively skewed index dataset (1-4) that results in the best fit (distribution pictured below code). Database ...
- 231
0 votes
0 answers
80 views
Train a neural network to predict the false positive rate of a segmentation model
I am trying to train a neural network to infer the false positive rate of an image segmentation model on the basis of the input image and the threshold. To do so, I am considering a dataset organized ...
