Add model error calculation

iandioch commented 8 years ago

Connects to #187

This is a standalone tool to get data from the database and figure out the error of the machine learning model for a particular user.

It's important to know the error of a model, so that when a change is made to our method, we can compare it objectively with the old method.

SailSlick commented 8 years ago

Any chance you could make a script so that it will run without having to set the python env yourself? Also what about writing it to a file so we can store results for presentation purposes?

devoxel commented 8 years ago

I don't those are big issues. Setting the python env is just:

source /home/python/bin/activate

Is it necessary to have a script to do it? It seems more complex than just doing it yerself :stuck_out_tongue:

#!/bin/bash
source /home/python/bin/activate
python /vagrant/visualisations/model_error/model_error.py "$@"

And writing it to a file is just:

python model_error.py username > out.txt

CianLR commented 8 years ago

Deffo document that if it's not there :+1: On 27 Apr 2016 11:53 a.m., "Aaron Delaney" notifications@github.com wrote:

I don't those are big issues. Setting the python env is just:

source /home/python/bin/activate

Is it necessary to have a script to do it? It seems more complex than just doing it yerself [image: :stuck_out_tongue:]

!/bin/bash

source /home/python/bin/activate python /vagrant/visualisations/model_error/model_error.py "$@"

And writing it to a file is just:

python model_error.py > out.txt

— You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub https://github.com/CPSSD/feedlark/pull/264#issuecomment-215048449

m1cr0man commented 8 years ago

Tested by me, reviewed by Cian.

CPSSD / feedlark

Add model error calculation #264

!/bin/bash