Documentation
Access all resources and documentation or contact us for additional assistance.
Before you start
Table Requirements
In order to run Brainome on your data, we require the following:
- Labeled Data
- CSV Format (RFC4180)
- Either do not include any columns with unique identifiers or use the -ignorecolumns option when executing Brainome
- For multi-class, provide a minimum of 100 data points per class
We also recommend the following practices:
- Having a unique header for each column; making the headers the first row of of your table
- Placing the output column at the far right (last column); otherwise, you will have to explicitly tell the system which column is the output
Important note: it is quite common to have one column be some kind of unique identifier for the row of data (typically alpha-numeric). If you do include this kind of column in your data, be sure to tell the system to ignore this column! We cover how to do this below.
Steps
1. Check that the Brainome tool is installed in the proper directory on your local machine:
which brainome
/usr/local/bin/brainome
If it is not, please refer to the Brainome PIP Installation or the Brainome Docker Installation
2. Verify the version by running a simple command:
brainome -version
brainome v1.002
3. Set up a directory to hold your training dataset:
pwd
/Users/user1/Documents/Brainome/Experiment/Experiment_data
ls -l
total 255560
-rw-r–r–@ 1 user1 staff 22115 Jul 26 14:07 experiment_test.csv
-rw-r–r–@ 1 user1 staff 22191 Jul 26 14:07 experiment_train.csv
For our purposes we use the Kaggle Titanic data, which is available at:
https://download.brainome.ai/data/public/titanic_train.csv
https://download.brainome.ai/data/public/titanic_validate.csv
https://download.brainome.ai/data/public/titanic_predict.csv