# TOM: Machine Learning-Based Test Results Analysis
This folder is created for TOM, a performance dataset processing tool, by Alassane Samba <alassane.samba@orange.com>
see https://wiki.opnfv.org/display/testing/R+post-processing+of+the+Yardstick+results

## deploy the TOM API server
Rscript tomRun.R

## load file
curl <your_local_ip>:8000/read?file=~/TOM/examples/test_opnfv_yardstick_tc012.csv

## analyze
curl <your_local_ip>:8000/analyze?input=deploy_scenario:version:pod_name:runner_id&output=bandwidth.MBps.

## view bivariate correlation between the inputs and the output
curl <your_local_ip>:8000/correlation

## view the analysis results in terms of output explaination rate obtained by combining inputs
curl <your_local_ip>:8000/explain

## view the corresponding graphic
curl <your_local_ip>:8000/explainGraph > explainGraph.png

## compare different PODs (boxplots)
curl <your_local_ip>:8000/comparGraph?input=pod_name&output=bandwidth.MBps.&limit=10 > comparGraph.png

## compare different configurations described by pod_name and deploy_scenario (boxplots)
curl <your_local_ip>:8000/comparGraph?input=pod_name:deploy_scenario&output=bandwidth.MBps.&limit=20 > comparGraph2.png