search

covid19india / deep-dive

This repo can contain all the analysis, machine learning work done using the patient data or the external data.
MIT License
31 stars 46 forks source link

Prediction of patients who needs to be admitted to ICU #3

Open gautam1858 opened 4 years ago

gautam1858 commented 4 years ago

Model that predicts whether a patient needs to be admitted to ICU

sumitkumarju commented 4 years ago

Sir I am working on this issue. i need some help- sir what are the units of the different columns in the patients.xsls. Also, it would be nice to get values for each category for normal patients which can be used to fill Nan values in the data frame as filling NAN values with column.mean() will hamper the model predictions

gautam1858 commented 4 years ago

You don't have to know units for the prediciton. (its unnecessary for example Hemoglobin unit is grams per deciliter )

sumitkumarju commented 4 years ago

Actaully i needed them to get the normal values for each category which further can be used as the default values to fill Nan values in the dataframe.

sumitkumarju commented 4 years ago

should i replace the nan values with the mean value for each column?

gautam1858 commented 4 years ago

Start by removing rows containing nan values, taking mean wouldn’t be a good idea

sumitkumarju commented 4 years ago

that leaves with zero rows in the dataframe

gautam1858 commented 4 years ago

this could help you https://www.kaggle.com/lucianacantarelli/task-1-prediction-covid-19

swapnanilsharma commented 4 years ago

I found most of the data are not updated. It seems ICMR is no more publishing "ICMRTestingDetails.csv" data anymore. Could you please help me with some authentic source for testing related data if you know any. Also please help me with updated "patients.xlsx" data or the source of the same.

Thank you