matthieukomorowski/AI_Clinician

About the generation of m*62 table

Opened this issue · 9 comments

bywbilly commented

Hi,

Thanks for your excellent work and code. I am wondering how to produce the m*62 table form the MIMIC III DB. Can you give me some ideas on producing that table?

Thanks!

matthieukomorowski commented
bywbilly commented

@matthieukomorowski
Thanks for updating your code. I want to extract the data from the mimiciii db. However, after running make-concept.sql in mimic_code, I still cannot extract Demographics data parts of your code.

Do you have any idea about that? Or could you just share me with your demog.csv, which would be very helpful.

moberst commented

@bywbilly In case it's helpful, I just ran the relevant SQL file directly to create a materialized view (elixhauser_quan.sql, I believe). I had to modify the code in AIClinician_Data_extract_MIMIC3_140219.ipynb slightly, because the materialized view was created in the mimiciii namespace instead of the public namespace, but otherwise it seemed to work

  • 👍1
bywbilly commented

@moberst Thank you very much! This actually works for me!

bywbilly commented

@matthieukomorowski @moberst Sorry to bother you again.
When I am running the AIClinician_sepsis3_def_160219.m, I got an error of this lineinputpreadm=table2array(readtable('D:/exportdir/preadm_fluid.csv')); because the first coloum is double while the second is cell. Then I change it to inputpreadm=table2cell(readtable('D:/exportdir/preadm_fluid.csv'));

And I still got an out of memory error, I ran the code on a 32GB Ubuntu. I am wondering how much memory will it cost?

Thanks!

moberst commented

Hey @bywbilly I made a new issue, since it sounds like your problem with preadm_fluid.csv is distinct from the original problem you mentioned. I'll respond to you in the new issue, I don't think that memory should be an issue

paulrich1234 commented

Hi,

Thanks for your excellent work and code. I am wondering how to produce the m*62 table form the MIMIC III DB. Can you give me some ideas on producing that table?

Thanks!

hello @bywbilly have you got the correct original data for training ,and i have run AIClinician_Data_extract_MIMIC3_140219.ipynb AIClinician_sepsis3_def_160219.m and AIClinician_mimic3_dataset_160219.m and i saved the MIMICtable = reformat4t; as csv file .but it seems there are some values difference https://github.com/paulrich1234/sepsisrl/blob/master/preprocessing/preprocess_data.ipynb in Out[12]:

could you tell me how to get it if you have got the correct one ? thank you in advance

Abdulk084 commented

Hi, I have created almost all the csv files but struggling to produce "demog.csv" in the notebook as given.
[https://github.com/Abdulk084/AI_Clinician/blob/master/AIClinician_Data_extract_MIMIC3_140219.ipynb] . using AWS Athena and had to change rest of the queries abit to produce csv files. Do you think that the provided notebook is compatible with AWS athena? Is there anyway I can get "demog.csv" so that I may step ahead to replicate the results mention in AI clinician paper.