CS 433 Machine Learning Project 2: Drug Treatment Classification and Lifespan Prediction of C. elegans using Machine Learning Techniques
(For those who are confused why everything is pushed in a single commit, this is because we have included confidential information of the drugs in the initial stages of the project and hence we have to remove the commit history for condifential concerns)
This repository contains the code of CS 433 Machine Learning Project 2. In this project, we utilise different machine learning models to disguish C. elegans worms treated with drugs from those not, and predict lifespan of these worms using early behavioral data, in specific, x-y coordinates of center of mass in the first 2 days.
Note: As the data set is small, we decided to include it in this repository. Users can directly run the code without extra steps to download the data
To get started, clone this repository:
git clone https://github.com/CS-433/ml-project-2-elegant-cda.git
cd ml-project-2-elegant-cda
and install the dependencies:
pip install -r requirements.txt --user
To run the code, run the command
python main.py
This command contains various arguments to be set by the users, for different tasks of the project, details of the arguments can be found by running
python main.py --help
For example, running the command
python main.py -m CNN -t clf -d 2 -s True -r True
will run the classification CNN training using data from drug 2, and will test it on a reproducible, random test split also from drug 2.
To modify the hyperparameters for a CNN model training, edit the values in the appropriate dictionary in config.py:
CNN_CLF_CONFIG: hyperparameters used for classification. Currently set as the hyperparameters values optimal for drug 2 training.
CNN_REG_CONFIG: hyperparameters used for regression. Currently set as the hyperparameters values optimal for drug 2 training.
GRID_SEARCH_REG: the set of hyperparameters tested in fine-tuning for the classification task.
GRID_SEARCH_CLF: the set of hyperparameters tested in fine-tuning for the regression task.
main.py contains the main file to be run for the project
data_loading.py contains functions to load the data into desired format
data_cleaning.py contains helper functions to clean the data, by filling nan values and remove data with excessive nan values
features.py generates features from the raw data
simple_models.py contains simple models from scikit-learn with default settings, and helper funcitons to train and evaluate the models on the project tasks.
neural_net.py is the model factory which generates the CNN model.
modelContainer.py is a model container object that contains a CNN model, and helper functions for training and evaluation for the project tasks.
utils.py contains other helper functions
data/Lifespan contains all the data for this project