Questions tagged [reproducible-research]
Research practice of making full experimental description, entire collected data, and all the data analysis scripts publicly available, so that the published results can be reproduced elsewhere.
85
questions
0
votes
0
answers
12
views
Does a newly constructed ML dataset need to have an official train-dev-test split? Should the test set be intentionally balanced?
I have constructed a novel ML (NLP) dataset for classification and labeled it with three classes. The dataset is rather small with about 700 examples, out of which the classes have about 400, 200, and ...
- 1
4
votes
1
answer
195
views
Fine-tuning of methods before main analysis: p-value hacking?
I am reviewing a paper for a journal. The authors propose a new classification scheme. They then try to assess the classification accuracy using some novel predictor data. Standard classification ...
- 93
0
votes
1
answer
20
views
How can Mass Spectrometry have high sensitivity but average reproducibility?
I am reading up on mass spec from this link: https://www.ebi.ac.uk/training/online/courses/metabolomics-introduction/designing-a-metabolomics-study/comparison-of-nmr-and-ms/
They state:
The two most ...
- 357
1
vote
0
answers
58
views
Python implementation of standard Expectation maximization algorithm to estimate position bias in LTR
Is there any Python implementation that would allow to use logs data of clicks and their ranking, and return and estimate of relative position bias as discussed in section ...
- 11
1
vote
0
answers
52
views
How to validate a clustering model on new data?
I was wondering how one would properly validate the reproducibility of a trained clustering model on a completely new set of data.
Imagine you are clustering a patient population in hospital A. You ...
- 11
1
vote
0
answers
30
views
How to ensure repeatability and reproducibility in a clinical study?
I have a 2 arm clinical study with topical anesthetic study group and placebo on the two arms. I measured the pain scores of patients with a pain scale. It is not an experimental study.In the method ...
- 71
0
votes
0
answers
34
views
A/B testing when there is a feedback loop
I am experimenting with changing label value (target) for a model that we have in production. We used to cap the target variable, and my new model will release the cap.
The main point about our ...
- 289
1
vote
0
answers
58
views
Can I use both a Kolmogorov–Smirnov test and a T-test to compare two distributions?
I have two distributions (N ~ 300, μ = 1.89 and σ = 0.97) and (N ~ 40, μ = 1.74 and σ = 1.18).
A ks_2samp returns statistic = 0.2 and p-value = 0.057. Therefore, I cannot reject the null hypothesis ...
- 41
3
votes
2
answers
104
views
Could we estimate replicability of empirical research with conformal predictions?
A review article Threats of a Replication Crisis in Empirical Computer Science reviews reproducibility issues. The authors present distinctions among repeatability, replicability, and reproducibility. ...
- 2,412
0
votes
0
answers
286
views
Help interpreting event study methodology of a famous research article
I wonder if someone out there familiar with reading research articles, especially finance research can help me interpret which event study methodology the author uses in her famous research article. ...
0
votes
1
answer
26
views
Gauge R&R analysis - Preicision of a system
I never attempted to do a GR&R analysis before so I'm sharing an overview of what I would like to attempt to do and hopefully I would get some insight/advice/reference on which approach I should ...
- 101
2
votes
0
answers
47
views
Understanding a simple post-hoc power analysis of effect size
I am trying to understand a relatively straightforward paragraph found in "Gun Shows and Gun Violence: Fatally Flawed Study Yields Misleading Results" by Wintermute et al that appeared in ...
- 21
2
votes
1
answer
211
views
Help in re-producing the calculation of the standardised difference for a multinomial variable
I am interested in calculating the standardised difference for a multinomial variable. In R there is a package called tableone that produces such calculations (https://cran.r-project.org/web/packages/...
- 175
1
vote
2
answers
403
views
Random number generator seed in R
I have a question about the random number generator seed number in R.
Recently, I am trying to solve the exercises in the book named "An Introduction to Statistical Learning".
When I ...
- 755
0
votes
1
answer
39
views
Best Practices in Sampling Large Datasets
I am working on a research paper concerning developing a CNN model for a multi-class classification on images. I have a large dataset consisting of 3 classes summing up to 20000 images. Class 1 has ...
- 51
3
votes
0
answers
34
views
To what degree does reproducibility of research differ across disciplines?
Have there been any statistics meta studies that compare the reproducibility of results, as well as the willingness of researchers to provide reproducible data, analysis, and code, in their studies ...
- 301
1
vote
1
answer
2k
views
Data science project examples based on Cookie Cutter Data science template [closed]
Interested by production ready data science project do you have any public github reference concrete implementation?
As an example i found this project template structure https://drivendata.github.io/...
- 31
0
votes
1
answer
54
views
What is the right machine learning research methodology to compare two neural network approaches against eachother?
I am new to machine learning research, and I have a general question regarding how to compare different models on the same data set. How does a top researcher do this, and what is expected from a ...
- 129
2
votes
0
answers
186
views
Ablation Study training until when?
Given that deep learning models can take weeks to train, having to train 20 different models becomes infeasible when performing an ablation study. Are there some formally accepted common practices on ...
- 445
0
votes
1
answer
274
views
Predict values from complex Rjags model
It's the first time I'm working with R2Jags, MCM chains and Bayesian models and I'm having trouble to compute the predicted values for my model.
The model is based on research by Hallmann et al. 2017, ...
- 1
1
vote
0
answers
20
views
Reproducibility Between Remote and Local Computer [closed]
I am currently using R, RStudio, and a remote computer (using an R script) to run the same ...
- 233
3
votes
1
answer
44
views
Appropriate model for amount of statistical errors in articles
I recently started my PhD and I am currently working on a project about finding statistical reporting errors. Our work is similar to Nuijten et al. (2016) only for economics.
So, I have a database ...
- 171
0
votes
0
answers
191
views
Failing to replicate the signal-to-noise ratio from original lasso paper
I'm referencing example 1 under the simulations section from Tibshirani.
In short there's a model in the form
$$
y = \beta^Tx + \sigma \epsilon,
$$
where $\beta = (3, 1.5, 0, 0, 2, 0, 0, 0)^T$ and $\...
- 1
0
votes
1
answer
307
views
Random number generator independence
I have a naïve question about random number generators.
Say I want to generate a set of random variables from different exponential distributions,$$X_i \sim \mathcal {Exp}(\lambda_i)$$ Assume the$X_i$'...
- 87
1
vote
1
answer
147
views
Pearson's Correlation between samples after PCA dimensionality reduction
This question is closely related to [Is it acceptable to reverse a sign of a principal component score?.
I have a matrix A with Samples in row and Features in column. I apply PCA on A and get a ...
- 11
0
votes
0
answers
25
views
Research on neural networks mimicking behavior
My question is regarding neural networks mimicking noise specifically.
The idea.
I have clinically clean image generated by a ray tracer and a neural network (type does not matter) that is mimicking ...
- 101
1
vote
0
answers
34
views
How to determine statistical significance of physics dataset?
I am working on a physics paper, which looks at the dependence of resistivity on physical temperature. I obtained data showing a linear trend of the form
$$\rho(T)=\rho_0[1+\alpha(T-T_0)]$$
Based on ...
- 111
0
votes
0
answers
31
views
Setting seeds despite repeated training of CNNs?
I would like to compare the classification performance (like accuracy, precision, recall etc.) of different CNN architectures. I'm using Google Colab (GPU support), Tensorflow and Keras.
Since it is ...
- 37
0
votes
1
answer
65
views
How to compare the performance of different SVMs and CNNs?
I'm a beginner in machine learning und I have a problem to find the best way to compare the performance (accuracy) of different SVMs and CNNs (Jupyter Notebook)?
The CNNs I train in Google Colab with ...
- 37
0
votes
1
answer
965
views
How to compare CNN models with non-reproducible results?
I try to compare different CNN models. I use Keras and for training, I use a GPU, Google Colab with Tensorflow backend.
Unfortunately I'm not able to create the same initial conditions for the CNNs (...
- 37
0
votes
0
answers
156
views
Does a high negative concordance imply reproducible features
Does a high negative(-1) Lin's concordance correlation coefficient(Lins CCC) imply reproducible features ?
I am calculating Lin's CCC of radiomic features between two different CT scan settings , ...
2
votes
1
answer
150
views
Is most published research indeed false?
I have come across (Ioannidis, 2005) which explains several reasons (mainly statistics-related, that's why I post this question here) to justify the claim that most published research is indeed false. ...
- 2,500
2
votes
1
answer
274
views
Reproducible numbers in Keras/TensorFlow
Every time I run a Keras/TensorFlow code gives different results. Can someone suggest how to get reproducible numbers?
- 658
9
votes
1
answer
520
views
Why do people use PCA when it has so many issues?
(This is a soft question) Recently I'm learning Principal Component Analysis, and it appears to have a lot of issues:
You have to transform the data to roughly the same scale before applying PCA, but ...
- 1,677
1
vote
0
answers
43
views
Replicating a wrong analysis
A study measures the performance of patients and controls in two tasks.
In task 1, a t-test reveals no significant differences between patients and controls. In task 2, a t-test reveals significant ...
- 381
4
votes
1
answer
158
views
Why replication studies use two-tailed tests?
I have been searching replication studies and I found that all of them use two-tailed tests. For example:
Open Science Collaboration. (2015). Estimating the reproducibility of psychological science....
- 381
7
votes
0
answers
70
views
Has the reproducibility crisis affected confidence intervals as well?
The reproducibility crisis has given many pause over the value (?) of $p$-values to measure the relevance of statistical findings. Given the interpretation of a $p$-value and some knowledge of ...
- 56.9k
0
votes
1
answer
223
views
What is a good machine learning academic research workflow/tools? [closed]
I'm interested in ML research and want to get some insights into how its done in practice.
In particular:
What kind of data are you dealing with?
How are you dealing with hyper-param optimization?
...
0
votes
0
answers
32
views
Use Deep learning or RL to maximize long term hit in caching file?
Suppose I have F video files with different file size and with different popularity.
I have a memory cache of fixed size. What I have to do is following:
I want to cache n files (n < F) in ...
- 692
6
votes
2
answers
10k
views
Choosing the "Correct" Seed for Reproducible Research/Results
I have created this post to spark a discussion around an idea I had regarding the choice of the seed number used to replicate the results of a statistical model. Here is some background on how I came ...
- 71
2
votes
0
answers
80
views
Put Together Results from GLM Gamma Models
I have a set of healthcare data in which I used GLM Gamma to model the healthcare spending. Then I am trying to put together the results in a style that is similar to this: http://onlinelibrary.wiley....
- 31
6
votes
1
answer
487
views
"report the statistical results of the analysis without the covariate"
This is from Psychological Science
22(11), p.1363:
If an analysis includes a covariate, authors must report the statistical results of the analysis without the covariate.
Reporting covariate-...
- 171
0
votes
1
answer
601
views
When to calculate Cronbach's alpha for a paper?
I am doing my undergraduate thesis, and I wonder if I need to do the Cronbach's alpha from my data set, or if I should use the Cronbach's alpha from the commercial questionnaire I will use for this ...
0
votes
1
answer
257
views
I have all negatives in (some) of my folds in cross validation. Seeking best way to calculate AUC using R in this scenario?
Problem
The title explains my problem. I cannot calculate AUC as things currently stand because some of my data sets have only negatives. I have three possible solutions and I need input from more ...
- 145
1
vote
0
answers
107
views
Two samples t-test and ANOVA
B) When doing a two way ANOVA
After calculating the sum of squares column(between), the sum of squares error(within), and the sum of squares block.
1. Are there any other sum of squares?
2. What is ...
1
vote
1
answer
74
views
Simple questions about reproducibility
Suppose we made $m$ tests in 2 samples from the population. For simplicity, we compared $m$ traits (i.e. bmi, temperature, height, etc) for equality in means between 2 groups of people (i.e. Germans ...
- 1,701
1
vote
1
answer
135
views
How to prove two methods are "similar"?
We are testing two imaging machines. One of the machines has been validated numerous times. However, we built the new machine and we need to prove that it is similar/comparable to the validated ...
- 111
0
votes
0
answers
175
views
Calculate Cronbach's alpha for three observers
I have a dataset which contains 100 observations, and I asked three expert to describe each observation by one of the letters {A,B,C}. The following three tables shows a sample of the dataset with the ...
- 399
1
vote
1
answer
49
views
Evidence for data mining (specification search) in published results [duplicate]
Is there an established way to assess the prevalence of data mining (as in specification search, not in the machine learning sense) in academic publications?
I vaguely remember hearing something ...
- 327
6
votes
1
answer
370
views
Statistical significance, repeatability, and sample size (50 shades of grey)
Reading about the problems with significance I wondered how one would approach an experiment in the right way, considering that in practice the sample size cannot be arbitrarily large.
While some ...
- 61