This application is capable of building the model, save it for future purpose, without writing a single piece of code!
A special library, Streamlit
is used to develop the application's interface. The documentation can be found here
- The application can be found here.
- Upload the
.csv
file you wanted to build the model on. - Select the features/columns form the drop-down menu.
- Handle the missing data(NaN) using different strategies. (A warning is displayed if it cannot be added. Try another strategy)
- Enccode the columns for training. (One-Hot encoder)
- Split the data into training and dev/test sets. (The max. split is set to 0.3 i.e., the dataset is split in the ratio, 70/30)
- Normalise the data.
- Select the algorithm for predicting.
- Modify the hyperparameters on the sidebar for better results.
- Click
save
button to save the model for later use.
- Clone the repository from above or in the commad line use:
$ git clone https://github.com/Nitin1901/machine-learning-interface.git
- Change you current working directory.
$ cd machine-learning-interface
- Create a virtual environment(recommended) and activate.
$ python -m venv ml-intreface
$ ml-interface\Scripts\activate.bat
- Install the required packages from
requirements.txt
. You can manually install each package or use:
$ pip install -r requirements.txt
- Open
app.py
in a text editor and start making changes. - Run the app locally
$ streamlit run app.py
If you wish to contribute, fork
the repository, develop and create a pull request
.