[HN Gopher] Machine-learning improves the prediction of stroke r...
___________________________________________________________________
Machine-learning improves the prediction of stroke recovery
Author : rustoo
Score : 16 points
Date : 2021-07-11 15:05 UTC (7 hours ago)
(HTM) web link (actu.epfl.ch)
(TXT) w3m dump (actu.epfl.ch)
| light_hue_1 wrote:
| Papers like this need machine learning reviewers and even better,
| coauthors. I cannot believe that anyone with competence in ML
| reviewed this paper.
|
| For one thing, the figure that the linked article presents is
| very misleading. It is not the performance of the model! It's
| just the raw data. Patients by how well we think they might
| recover vs. how well they discovered. The goal of the model is to
| figure out how will recover well. Literally, nothing in that
| figure is an output of the model.
|
| > In case of stroke lesions affecting the performance of
| parcellation, the correspondence part of the lesion on the
| unaffected hemisphere was used and interpolated over the lesioned
| brain as a brain transplant and fed into the above-described
| structural imaging processing.
|
| The classifier will trivially discover that you are doing this,
| and it's almost certainly correlated with terrible outcomes. (PS:
| should be "corresponding" part) This alone calls into question
| all of the results and never should have gotten past anyone. Just
| because you can't tell as a human, doesn't mean the data
| manipulation isn't immediately obvious to even a simple model
| like an SVM.
|
| > By means of k-means clustering, patients were identified as
| fitters, i.e. revealing a proportional recovery, and non-fitters,
| i.e. patients lacking proportional recovery.
|
| Good god, why? Just set up a simple criteria and use it. This
| kind of strange arbitrary boundary in your data that comes from
| k-means can only hurt your model.
|
| Finally: The results are terrible. When cross-validating the
| model does well, 83% accuracy (87% precision). But on the
| external data the model is worthless. 60% accuracy, 53%
| precision.
|
| They provide no trivial baselines. They don't even list of the
| base rates for their dataset (I'm guessing that out of their main
| dataset the classifier is just operating at chance at 60%
| accuracy / 53% precision is just the base rate).
|
| Their data seems to be horrific. Half the patients in the 2nd
| dataset don't have data for both time points. How can they be
| assigned to groups that fit or don't fit without this basic data?
| They don't even describe results on their 3rd dataset.
|
| They don't bother to compute a confidence interval, which will be
| huge because they have so little data.
|
| Seriously. Authors. If you are reading this, reach out to a
| machine learning person. If you don't know anyone, reach out to
| me, I'm a scientist at a major US university. Papers like this
| just pollute our science and are a total waste of time on your
| part.
| rscho wrote:
| You're just not used to clinical research. A large part of the
| field output is no better than this. And this is only what the
| authors think will look good on paper, which is often quite far
| from what really happened during data collection and analysis.
| I should know, I'm a clinical researcher at a major european
| uni ;-) And due to peer pressure, my output is also no
| better... :-(
| AlecSchueler wrote:
| Would be curious to hear more about the systemic issues that
| lead to you knowingly outputting poor research. Could you
| perhaps blog about it anonymously?
| riahi wrote:
| Agreed. I try to accept all ML focused paper review requests in
| my disciplines, but modern academics do not reward you for the
| time you take as a reviewer, which then distracts me from
| working on my own projects.
|
| It's very much an uncompensated service to the community...
| rob_c wrote:
| Hope this doesn't mean that we hold back some treatment for those
| who would have bucked the trend and recovered better than
| expected... Very interesting if this can be used to tailor
| recovery and treatments in an adaptive way. :)
| karmicthreat wrote:
| Yea, this seems more like a tool insurance companies will use
| to deny coverage.
| deeviant wrote:
| What is the value of such a prediction?
___________________________________________________________________
(page generated 2021-07-11 23:01 UTC)