To what ever poor soul who has found this repo, welcome! Enclosed is a collection of ML algorithms I wrote during CSC 692 at SDSM&T.
Assuming that you have the repo cloned, we can start with installing the requirements.
Do this by running the following command:
pip install -r requirements.txt
Once that finishes, you should be able to utilize this ML package.
For every algorithm in the package, there are 3 functions that you will need to know.
On init of each of the algorithms there will usually be a couple of optional arguments that modify parameters of the model. Such as weight scales or epochs. To pass these optional arugments in, simpily declare the object like this:
model = Algorithm(scale=.1, epochs=1000)
And just like that, you have inistantiated the algorithm! For more detailed information on the initialization of each algorithm, please see the algorithms documentation or docstring.
Fitting the model is the most important step of the usage. Without fitting the model you cannot use it to predict/classify/etc.
In order to fit the models, each model requires datapoints and the expected label/result for each data point.
To fit the algorithm, use the following format:
model.fit(data_points, labels)
Once the model has been trained, the model can then be used. For information on how to use a specific model, see the model/algorithm's documentation.