.csv file

[HodaMemar]

Hi Can you explain about the .csv file? I could not find it. csv/pneumonia_admissions.csv

[markozeman]

Hi, all .csv files are made from public MIMIC II database, but are too big to upload here.

[HodaMemar]

Thanks Is it possible to show the format of these files? Only head of files with a few records from the csv will be sufficient so I can make it like that. Is there pneumonia_admissions.csv is admissions.csv that filter on special disease? On Mon, 6 Jul 2020, 16:33 Marko Zeman, notifications@github.com wrote:

Hi, all .csv files are made from public MIMIC II database, but are too big to upload here.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/markozeman/KarmaLego/issues/1#issuecomment-654192103, or unsubscribe https://github.com/notifications/unsubscribe-auth/AP4V5FNTVQOFTVOLL3SGFDDR2G4SLANCNFSM4OQWVICQ .

[markozeman]

pneumonia_admissions.csv is a subset of all_admissions.csv, filtered by only one diagnosis.

pneumonia_admissions.csv format: patient_id | startdate | enddate | drug | admittime | dischtime | diagnosis 68 | 2173-12-16 | 2173-12-16 | Diltiazem | 2173-12-15 | 2174-01-03 | PNEUMONIA

electrolytes.csv document copied: Electrolytes | occurrences Potassium Chloride | 192993 D5W | 142241 (5% raztopina dekstroze) 0.9% Sodium Chloride | 130147 NS | 129731 Iso-Osmotic Dextrose | 87005 Sodium Chloride 0.9% Flush | 83392 5% Dextrose | 73829 Calcium Gluconate | 52110 Bag | 25745 Vial | 24502 Dextrose 50% | 21797 D5 1/2NS | 16276 Sterile Water | 14100 0.9% Sodium Chloride (Mini Bag Plus) | 14048 Syringe | 13383 Iso-Osmotic Sodium Chloride | 11581 Chlorhexidine Gluconate 0.12% Oral Rinse | 11562 Multivitamins | 10682 1/2 NS | 10294 Dextrose 5% | 9621 Potassium Chl 20 mEq / 1000 mL D5 1/2 NS | 8870 Calcium Carbonate | 8246 NS (Mini Bag Plus) | 7956 D5W (EXCEL BAG) | 7182 Soln. | 7044 Soln | 6654 D10W | 6633 Sodium Chloride | 6550 NS (Glass Bottle) | 5894 Lactated Ringers | 5413 Syringe (Neonatal) D5W | 5060 Syringe (Neonatal) | 4524 Potassium Chloride (Powder) | 4260 Potassium Chl 20 mEq / 1000 mL NS | 4217 Potassium Chloride Replacement (Oncology) | 3658 Syringe (IV Room) | 3606 5% Dextrose (EXCEL BAG) | 3527 Send 500mg Vial | 3315 D5NS | 3164 SW | 72522

[HodaMemar]

Thanks. It will be very helpful

[markozeman]

You're welcome.

[HodaMemar]

I create a file for 3 patient and start from MIMIC_prescriptions2KarmaLego and uncomment in last line (write2json) but run time is very long. Could you please send one sample json file : pneumonia_entity_list_group_3.json

[markozeman]

It shouldn't be long for only 3 samples. You can use hand-made examples in file entities.py to understand working procedure.

[HodaMemar]

Thank you for your help I set : use_MIMIC = False remove_some_drugs = False

and I can use of hand-made examples in file entities.py

[HodaMemar]

Excuse me Sir I have a question about algorithm, The text of the reference article shows an example: three weeks of moderately high blood glucose,followed by five weeks of high-dose intermediate-acting insulin administration In this example, it can be assumed that there are two symbols: blood glucose and intermediate-acting insulin administration If the behavior of one of these symbols, such as blood glucose, changes in a zigzag pattern (low,high,low,high,...), is this considered in the karmalego algorithm? Can a symbol have intermittent intervals that are not necessarily of the same type?

https://pdfs.semanticscholar.org/a800/83f16631756d0865e13f679c2d5084df03ae.pdf

[markozeman]

If the behavior of one of these symbols, such as blood glucose, changes in a zigzag pattern (low,high,low,high,...), is this considered in the karmalego algorithm? Yes, blood glucose would be a symbol that can appear in different states we choose, for example low and high (we decide on the threshold). KarmaLego finds patterns in temporal data, for example it would find a pattern of [low, low, low] more often than pattern [high, low, high].

Can a symbol have intermittent intervals that are not necessarily of the same type? Yes. I believe it is meant to observe exactly that.