|
|
|
# EM-Bright random forest review
|
|
|
|
## Outlook of the code
|
|
|
|
Review work can be divided into three portion
|
|
|
|
1. Random forest implementation
|
|
|
|
2. Condor dag writer script
|
|
|
|
3. Implementation in em_bright.py
|
|
|
|
|
|
|
|
### Random forest implementation
|
|
|
|
Relevant lines : [random forest](https://git.ligo.org/sushant.sharma-chaudhary/em-bright-rf/-/blob/random/ligo/em_bright/utils.py#L286-312)
|
|
|
|
|
|
|
|
It takes the dataset form Deep's KNN framework, split them to train test (70% training and 30% testing), and then training set is passed onto RandomForest classifier. The **kwargs for the classifier is taken from the config file (./etc/config).
|
|
|
|
|
|
|
|
### Condor dag writer script
|
|
|
|
Script : [Dag Writer](https://git.ligo.org/sushant.sharma-chaudhary/em-bright-rf/-/blob/random/ligo/em_bright/dag_writer.py)
|
|
|
|
|
|
|
|
This code writes the workflow dag in order to submit to condor. It again take's Deep's common workflow scheme and separates for KNN and RandomForest classifier.
|
|
|
|
|
|
|
|
### Implementation in em_bright.py |
