Regression

Overview

The RegressionTracker class, inheriting from BaseTracker, is designed specifically for tracking regression models. It takes parameters needed for regression models, including variable names, hyperparameters for test/train split, and the model itself.

RegressionTracker

Bases: BaseTracker

Tracker object for regression models.

Source code in sientia_tracker/regression.py 
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
class RegressionTracker(BaseTracker):
    """
    Tracker object for regression models.
    """
    def __init__(self, tracking_uri, username: str = None, password: str = None):
        super().__init__(tracking_uri, username, password)

    def save_experiment(self,
                        model: Any,
                        model_name: str = "regr_model",
                        dataset_name: str = "data",
                        inputs: str | List[str] = "inputs",
                        target: str = 'target',
                        date_column: str = "date",
                        r2: float = 1.0, 
                        train_size: float = 0.8,
                        shuffle: bool = True
                        ) -> mlflow.ActiveRun:
        """
        Start a run in MLflow.

        params:
            model: Scikit-Learn Regression Model
            model_name: The name of the model as defined by user. Default: regr_model
            dataset_name: The name of the dataset. Default: data
            inputs: The name of the inputs. Default: inputs
            target: The name of the target. Default: target
            date_column: The name of the date column. Default: date
            r2: r2-score of the model. Default: 1
            train_size: Proportion of the data used to traind the model. Default: 0.8
            shuffle: Wheter or not the train data was shuffled. Default: True

        returns:
            active_run: The active run in MLflow
        """
        mlflow.end_run()


        print("Saving experiment", self.project_name)
        runs = mlflow.search_runs(experiment_names=[
            self.project_name], order_by=["start_time desc"])
        next_run_number = len(runs) + 1

        active_run = mlflow.start_run(run_name=f"{self.project_name}-{next_run_number}")
        mlflow.log_params({
            "Model": "Linear Regression",
            "Dataset": dataset_name,
            "Inputs": inputs,
            "Date Column": date_column,
            "Target": target,
            "Train Size": train_size,
            "Shuffle": shuffle,
        })

        mlflow.sklearn.log_model(model,model_name)
        mlflow.log_metrics({"r2":r2})

        return active_run

save_experiment(model, model_name='regr_model', dataset_name='data', inputs='inputs', target='target', date_column='date', r2=1.0, train_size=0.8, shuffle=True)

Start a run in MLflow.

Parameters:
  • model (Any) –

    Scikit-Learn Regression Model

  • model_name (str, default: 'regr_model' ) –

    The name of the model as defined by user. Default: regr_model

  • dataset_name (str, default: 'data' ) –

    The name of the dataset. Default: data

  • inputs (str | List[str], default: 'inputs' ) –

    The name of the inputs. Default: inputs

  • target (str, default: 'target' ) –

    The name of the target. Default: target

  • date_column (str, default: 'date' ) –

    The name of the date column. Default: date

  • r2 (float, default: 1.0 ) –

    r2-score of the model. Default: 1

  • train_size (float, default: 0.8 ) –

    Proportion of the data used to traind the model. Default: 0.8

  • shuffle (bool, default: True ) –

    Wheter or not the train data was shuffled. Default: True
Returns:
  • active_run( ActiveRun ) –

    The active run in MLflow
Source code in sientia_tracker/regression.py 
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
def save_experiment(self,
                    model: Any,
                    model_name: str = "regr_model",
                    dataset_name: str = "data",
                    inputs: str | List[str] = "inputs",
                    target: str = 'target',
                    date_column: str = "date",
                    r2: float = 1.0, 
                    train_size: float = 0.8,
                    shuffle: bool = True
                    ) -> mlflow.ActiveRun:
    """
    Start a run in MLflow.

    params:
        model: Scikit-Learn Regression Model
        model_name: The name of the model as defined by user. Default: regr_model
        dataset_name: The name of the dataset. Default: data
        inputs: The name of the inputs. Default: inputs
        target: The name of the target. Default: target
        date_column: The name of the date column. Default: date
        r2: r2-score of the model. Default: 1
        train_size: Proportion of the data used to traind the model. Default: 0.8
        shuffle: Wheter or not the train data was shuffled. Default: True

    returns:
        active_run: The active run in MLflow
    """
    mlflow.end_run()


    print("Saving experiment", self.project_name)
    runs = mlflow.search_runs(experiment_names=[
        self.project_name], order_by=["start_time desc"])
    next_run_number = len(runs) + 1

    active_run = mlflow.start_run(run_name=f"{self.project_name}-{next_run_number}")
    mlflow.log_params({
        "Model": "Linear Regression",
        "Dataset": dataset_name,
        "Inputs": inputs,
        "Date Column": date_column,
        "Target": target,
        "Train Size": train_size,
        "Shuffle": shuffle,
    })

    mlflow.sklearn.log_model(model,model_name)
    mlflow.log_metrics({"r2":r2})

    return active_run