Multilabel-timeseries-classification-with-LSTM
Multilabel-timeseries-classification-with-LSTM copied to clipboard
aqibsaeed
Testing Dataset
Could you provide a small training data sets even just for running the codes as the dataset is quite difficult and time-consuming to get.
Thx
Very much hope that you can provide a small training data set for training data .thanks
Thank you for your interest. Currently I don't have any open dataset available. I will definitely update the repository as I get one. It's on my to-do list :)
Can you tell me the format of the data ? The paper said the data consists of 10401 episodes, each a multivariate time series of 13 variables, Episodes vary in length from 12 hours to several months.I want to know that data.csv how to store the data . thank you .please
very thanks .but i want to know how you to store the 3 dimensional data [batch_size, time_steps, number_of_features] into the data.csv ? and the window_size is what ?
data.csv is a placeholder word for filename! The data in the file is not 3d, you have to parse your data using sliding window or related approach and save it in numpy array.
the one episodes is one batch ? and i have another question , in the code signal = data.ix[start:end]["<FEATURE COLUMN NAME>"] ,this is only one feature? , I try to use two feature ,it is wrong
Cleaned version of MIMIC-III is now available at: https://github.com/YerevaNN/mimic3-benchmarks
@aqibsaeed Thanks. Is this dataset same as the data used in this jupyter notebook example?
Unfortunately that dataset is not publicly available. But this cleaned version of MIMIC-III is similar to that dataset.
@aqibsaeed What is the difference between this cleaned version and the dataset used in this repo. It would be very nice if you can provide some example of it to reproduce the results of your implementation.
The dataset used in the paper, was from different hospital and not publicly available. MIMIC-III dataset is publicly available but is hard to work with, the cleaned version provides scripts to process the MIMIC-III for different problems. You have to download MIMIC-III yourself, I can't upload or provide any example of that due to restrictions.