Hi Solly,
I had a quick look at available data in " demgraphic training" and "walking activity training" and found some columns that should not be available in test set. For example in " demgraphic training" there is a column named "diagnosis year" also "healthcare provider" and in ""walking activity training" file the last column that is medTimepoint. Can you please let us know which part of data will not be available in test sets? Thanks! - Samad
Created by S J SAJA Can you please release the test dataset table? if the whole table is not available, can you please release such as 50 lines exactly as the format in the final test?
Can you please provide the code for the base-learners that you are going to test. For example, depending on the tree depth you use, we will have to decide on our side the number of features I provide. If it is a series of base-learners/parameters, can you please provide all as well as whether you shall use the mean/median or max performance of all?
thank you Dear @kishoresubu, I think your question is answered in the data description [wiki](#!Synapse:syn8717496/wiki/448355). Please let me know if some description is not clear. Can some one explain me the columns in the demographics data set ,walking data set and also the explanation for values in it. Why does the data set have these columns? Dear Samad:
In order to focus the challenge on accelerometer data we will supply a minimal set of covariate data. Specifically we will provide the following columns in test walking data:
**recordId**
**healthCode**
**createdOn**
**appVersion**
**phoneInfo**
**accel_walking_outboundjsonitems**
**deviceMotion_walking_outboundjsonitems**
**pedometer_walking_outboundjsonitems**
**accel_walking_returnjsonitems**
**deviceMotion_walking_returnjsonitems**
**pedometer_walking_returnjsonitems**
**accel_walking_restjsonitems**
**deviceMotion_walking_restjsonitems**
For the demographics, we will include covariates in our scoring mechanism, for example we know that age is associated to performance and will use this variable in the model building and testing but expect that features will not be built using different methodologies depending on these covariates. With that being said we will at the very last supply the following columns in the demographics:
**recordId**
**healthCode**
**createdOn**
**Date**
**phoneInfo**
But we are currently having an internal debate on whether we will also supply additional columns (age, gender, etc). We will get back to you as soon as we know the answer to this.
Best,
Larsson
Drop files to upload
Question about available data in test sets page is loading…