148 lines
5.1 KiB
Markdown
148 lines
5.1 KiB
Markdown
# <img src="logo.png" alt="logo" width="50"/> Platform
|
|
|
|
|
|
[](<https://opensource.org/licenses/MIT>)
|
|
|
|
|
|
Platform to run Bayesian Networks and Machine Learning Classifiers experiments.
|
|
|
|
## 0. Setup
|
|
|
|
Before compiling Platform.
|
|
|
|
### Miniconda
|
|
|
|
To be able to run Python Classifiers such as STree, ODTE, SVC, etc. it is needed to install Miniconda. To do so, download the installer from [Miniconda](https://docs.conda.io/en/latest/miniconda.html) and run it. It is recommended to install it in the home folder.
|
|
|
|
In Linux sometimes the library libstdc++ is mistaken from the miniconda installation and produces the next message when running the b_xxxx executables:
|
|
|
|
```bash
|
|
libstdc++.so.6: version `GLIBCXX_3.4.32' not found (required by b_xxxx)
|
|
```
|
|
|
|
The solution is to erase the libstdc++ library from the miniconda installation:
|
|
|
|
### MPI
|
|
|
|
In Linux just install openmpi & openmpi-devel packages. Only if cmake can't find openmpi installation (like in Oracle Linux) set the following variable:
|
|
|
|
```bash
|
|
export MPI_HOME="/usr/lib64/openmpi"
|
|
```
|
|
|
|
In Mac OS X, install mpich with brew and if cmake doesn't find it, edit mpicxx wrapper to remove the ",-commons,use_dylibs" from final_ldflags
|
|
|
|
```bash
|
|
vi /opt/homebrew/bin/mpicx
|
|
```
|
|
|
|
### boost library
|
|
|
|
[Getting Started](<https://www.boost.org/doc/libs/1_83_0/more/getting_started/index.html>)
|
|
|
|
The best option is install the packages that the Linux distribution have in its repository. If this is the case:
|
|
|
|
```bash
|
|
sudo dnf install boost-devel
|
|
```
|
|
|
|
If this is not possible and the compressed packaged is installed, the following environment variable has to be set pointing to the folder where it was unzipped to:
|
|
|
|
```bash
|
|
export BOOST_ROOT=/path/to/library/
|
|
```
|
|
|
|
In some cases, it is needed to build the library, to do so:
|
|
|
|
```bash
|
|
cd /path/to/library
|
|
mkdir own
|
|
./bootstrap.sh --prefix=/path/to/library/own
|
|
./b2 install
|
|
export BOOST_ROOT=/path/to/library/own/
|
|
```
|
|
|
|
Don't forget to add the export BOOST_ROOT statement to .bashrc or wherever it is meant to be.
|
|
|
|
### libxlswriter
|
|
|
|
```bash
|
|
cd lib/libxlsxwriter
|
|
make
|
|
make install DESTDIR=/home/rmontanana/Code PREFIX=
|
|
```
|
|
|
|
### Release
|
|
|
|
```bash
|
|
make release
|
|
```
|
|
|
|
### Debug & Tests
|
|
|
|
```bash
|
|
make debug
|
|
```
|
|
|
|
### Configuration
|
|
|
|
The configuration file is named .env and it should be located in the folder where the experiments should be run. In the root folder of the project there is a file named .env.example that can be used as a template.
|
|
|
|
## 1. Commands
|
|
|
|
### b_list
|
|
|
|
List all the datasets and its properties. The datasets are located in the _datasets_ folder under the experiments root folder. A special file called all.txt with the names of the datasets has to be created. This all file is built wih lines of the form:
|
|
<name>,<class_name>,<real_features>
|
|
|
|
where <real_features> can be either the word _all_ or a list of numbers separated by commas, i.e. [0,3,6,7]
|
|
|
|
### b_grid
|
|
|
|
Run a grid search over the parameters of the classifiers. The parameters are defined in the file _grid.txt_ located in the grid folder of the experiments. The file has to be created with the following format:
|
|
|
|
```json
|
|
{
|
|
"all": [
|
|
<set of hyperparams>, ...
|
|
],
|
|
"<dataset_name>": [
|
|
<specific set of hyperparams for <dataset_name>>, ...
|
|
],
|
|
}
|
|
```
|
|
|
|
The file has to be named _grid_<model_name>_input.json_
|
|
|
|
As a result it builds a file named _grid_<model_name>_output.json_ with the results of the grid search.
|
|
|
|
The computation is done in parallel using MPI.
|
|
|
|
|
|
|
|
### b_main
|
|
|
|
Run the main experiment. There are several hyperparameters that can set in command line:
|
|
|
|
- -d, -\-dataset <dataset_name> : Name of the dataset to run the experiment with. If no dataset is specificied the experiment will run with all the datasets in the all.txt file.
|
|
- -m, -\-model <classifier_name> : Name of the classifier to run the experiment with (i.e. BoostAODE, TAN, Odte, etc.).
|
|
- -\-discretize: Discretize the dataset before running the experiment.
|
|
- -\-stratified: Use stratified cross validation.
|
|
- -\-folds <folds>: Number of folds for cross validation (optional, default value is in .env file).
|
|
- -s, -\-seeds <seed>: Seeds for the random number generator (optional, default values are in .env file).
|
|
- -\-no-train-score: Do not calculate the train score (optional), this is useful when the dataset is big and the training score is not needed.
|
|
- -\-hyperparameters <hyperparameters>: Hyperparameters for the experiment in json format.
|
|
- -\-hyper-file <hyperparameters_file>: File with the hyperparameters for the experiment in json format. This file uses the output format of the b_grid command.
|
|
- -\-title <title_text>: Title of the experiment (optional if only one dataset is specificied).
|
|
- -\-quiet: Don't display detailed progress and result of the experiment.
|
|
|
|
### b_manage
|
|
|
|
Manage the results of the experiments.
|
|
|
|
### b_best
|
|
|
|
Get and optionally compare the best results of the experiments. The results can be stored in an MS Excel file.
|
|
|
|
|