MooLML

■ In case of upload, the filename have to be in English and any of space is NOT allowed. Otherwise it might occur errors or malfunctioning by character encoding while modeling process.

■ Upload data must be Comma-Seperated Values files(*.csv) consisting of characters and numbers.
■ Write the feature name(Field Name) in English and configure the last field into a target column.
■ The name of features can be character and number, but special characters such as ?, <, > is not allowed.

■ To upload the input data, the file have to be form of *.csv format which is Comma-Seperated Values. User can easily drag and drop file from File Explorer or select a file with Add Files button.

■ Currently, 2 modeling methods are available. You can select specific modeling method.

■ Percentage Test Split: User defines the amount of test result by percentage(%).
■ Use Training Set: Train with 100% input data and test it with random data in 20% range.
■ Supplied Test Set: Both input data and test data have to be uploaded. WaterML conducts test by second file on upload form.
■ Cross-Validation: It validate the test result as many as Cross-Validation option given by user.

■ During process, user can easily check how far the process is done.
① Process WaterML engine do.
② Statistics of input data.
③ The epoch processes with loss and accuracy.

■ On the calculation result page, user can see the modeling & test result intuitively.
■ Visualization: Machine Learning result with accuracy, Confusion Matrix(Classification), Scatter(Regression), Result file
■ Download: Machine Learning Result, Models(*.h5, *.pkl), Confusion Matrix(Classification), Scatter(Regression), Result file
■ Test Code: A Python-based test code is provided for easy applying with How-To-Run.

■ Sample test code with instruction is provided for applying.