Working with data

This example explains how to load your data.

Fedot provides specific interface for operation with data. Fedot uses it’s own data object notation (InputData). It contains index, features and target for each sample. You can create it from file using InputData.from_csv() method. You need to provide Task object with type of task you want to solve. Here examples of for tabular data:

from fedot.core.data.data import InputData
data_path = 'path_to_data'

data = InputData.from_csv(data_path,
target_columns='target',
task=Task(TaskTypesEnum.classification)) # or regression

Note

There are 3 possible values for TaskType: * TaskTypesEnum.classification * TaskTypesEnum.regression * TaskTypesEnum.ts_forecasting

Note

You can provide several target columns (For regression task).Then Fedot will recognise it as multiregression task supported natively.

You also can create InputData from pandas DataFrame:

from fedot.core.data.data import InputData


data = InputData.from_dataframe(features_df,
                                target_df,
                                task=Task(TaskTypesEnum.classification)) # or regression

or from numpy array:

from fedot.core.data.data import InputData


data = InputData.from_numpy(features_array,
                            target_array,
                            task=Task(TaskTypesEnum.classification)) # or regression

After you can split data on train/test set:

train, test = train_test_data_setup(data)

and pass it to the model:

model = Fedot(...)
model.fit(train)
model.predict(test)

For time series forecasting problem there is a little bit different approach for data initialization. Firstly you need to create a Task object:

from fedot.core.repository.tasks import Task, TaskTypesEnum, TsForecastingParams
# specify the task and the forecast length (required depth of forecast)
task = Task(TaskTypesEnum.ts_forecasting,
            TsForecastingParams(forecast_length=your_forecast_length))

After that you can use Input_data.from_csv_series

train_input = InputData.from_csv_time_series(task=task,
                                             file_path='time_series.csv',
                                             delimiter=',',
                                             target_column='value')

But you also can create InputData from numpy :

train_input = InputData.from_numpy_time_series(series,
                                               task=task)

After you can split data on train/test set (test set will contain last N values of the series by default):

train, test = train_test_data_setup(data)

and pass it to the model:

model = Fedot(...)
model.fit(train)
model.forecast()

Thus, this example shows how to operate with data in Fedot.