View Jupyter notebook on the GitHub.

Outliers

This notebook contains the simple examples of outliers handling using ETNA library.

Table of contents

• Loading dataset

• Point outliers

  • Median method

  • Density method

  • Prediction interval method

  • Histogram method

• Interactive visualization

• Outliers imputation

!pip install "etna[prophet]" -q

import warnings

warnings.filterwarnings("ignore")

1. Loading dataset

import pandas as pd

from etna.datasets import TSDataset
from etna.metrics import MAE
from etna.metrics import MSE
from etna.metrics import SMAPE
from etna.models import MovingAverageModel
from etna.models import ProphetModel
from etna.pipeline import Pipeline

Let’s load and look at the dataset

df = pd.read_csv("data/example_dataset.csv")
ts = TSDataset(df, freq="D")
ts.head(5)

segment	segment_a	segment_b	segment_c	segment_d
feature	target	target	target	target
timestamp
2019-01-01	170	102	92	238
2019-01-02	243	123	107	358
2019-01-03	267	130	103	366
2019-01-04	287	138	103	385
2019-01-05	279	137	104	384

As you can see from the plots, all the time series contain outliers - abnormal spikes on the plot.

ts.plot()

In our library, we provide methods for point outliers detection, visualization and imputation. In the sections below, you will find an overview of the outliers handling tools.

2. Point outliers

Point outliers are stand alone abnormal spikes on the plot. Our library contains four methods for their detection based on different principles. The choice of the method depends on the dataset.

from etna.analysis import get_anomalies_density
from etna.analysis import get_anomalies_hist
from etna.analysis import get_anomalies_median
from etna.analysis import get_anomalies_prediction_interval
from etna.analysis import plot_anomalies

Note: you can specify the column in which you want search for anomalies using the in_column argument.

2.1 Median method

To obtain the point outliers using the median method we need to specify the window for fitting the median model.

anomaly_dict = get_anomalies_median(ts, window_size=100)
plot_anomalies(ts, anomaly_dict)

2.2 Density method

It is a distance-based method for outliers detection.

anomaly_dict = get_anomalies_density(ts, window_size=18, distance_coef=1, n_neighbors=4)
plot_anomalies(ts, anomaly_dict)

2.3 Prediction interval method

It is a model-based method for outliers detection. Outliers here are all points out of the prediction interval predicted with the model.

Note: method is now available only for ProphetModel and SARIMAXModel.

anomaly_dict = get_anomalies_prediction_interval(ts, model=ProphetModel, interval_width=0.95)
plot_anomalies(ts, anomaly_dict)

15:33:17 - cmdstanpy - INFO - Chain [1] start processing
15:33:17 - cmdstanpy - INFO - Chain [1] done processing
15:33:17 - cmdstanpy - INFO - Chain [1] start processing
15:33:17 - cmdstanpy - INFO - Chain [1] done processing
15:33:17 - cmdstanpy - INFO - Chain [1] start processing
15:33:17 - cmdstanpy - INFO - Chain [1] done processing
15:33:17 - cmdstanpy - INFO - Chain [1] start processing
15:33:17 - cmdstanpy - INFO - Chain [1] done processing

2.4 Histogram method

This method detects outliers in time series using histogram model. Outliers here are all points that, when removed, result in a histogram with a lower approximation error.

Note: method might work sufficiently slow.

anomaly_dict = get_anomalies_hist(ts, bins_number=10)
plot_anomalies(ts, anomaly_dict)

3. Interactive visualization

The performance of outliers detection methods significantly depends on their hyperparameters values. To select the best parameters’ configuration for the chosen method, you can use our interactive visualization tool.

from etna.analysis import plot_anomalies_interactive

You only need to specify segment, the outliers detection method and it’s parameters grid in format (min, max, step) for each parameter you want to control.

segment = "segment_c"
method = get_anomalies_median
params_bounds = {"window_size": (40, 70, 1), "alpha": (0.1, 4, 0.25)}

In some cases there might be troubles with this visualisation in Jupyter notebook, try to use !jupyter nbextension enable --py widgetsnbextension

plot_anomalies_interactive(ts=ts, segment=segment, method=method, params_bounds=params_bounds)

Let’s assume that the best parameters are:

best_params = {"window_size": 60, "alpha": 2.35}

4. Outliers imputation

The previous sections are focused on the outliers detection as the part of the EDA, however outliers imputation might be also one of the step in the pipeline of your forecasting model. Let’s explore a simple example of how to build such pipeline.

from etna.analysis import plot_imputation
from etna.transforms import MedianOutliersTransform
from etna.transforms import TimeSeriesImputerTransform

Segment with outliers:

df = ts[:, segment, :]
ts = TSDataset(df, freq="D")
ts.plot()

Outliers imputation process consists of two steps:

1. Replace the outliers, detected by the specific method, with NaNs using the instance of the corresponding outliers transform.

2. Impute NaNs using the TimeSeriesImputerTransform.

Before the imputation step you firstly need to choose the appropriate imputation method and you can do it using our EDA method for imputation visualisation - plot_imputation.

# Impute outliers with NaNs
outliers_remover = MedianOutliersTransform(in_column="target", **best_params)

ts.fit_transform([outliers_remover])
ts.plot()

Let’s impute outliers detected by median method using the running_mean imputation strategy.

# Impute NaNs using the specified strategy
outliers_imputer = TimeSeriesImputerTransform(in_column="target", strategy="running_mean", window=30)

Let’s look at the points that we replaced with TimeSeriesImputerTransform using plot_imputation.

plot_imputation(imputer=outliers_imputer, ts=ts)

ts.fit_transform([outliers_imputer])

Now you can fit the model on the cleaned dataset, which might improve the forecasting quality. Let’s compare the forecasts of the model fitted on cleaned and uncleaned train timeseries.

def get_metrics(forecast, test):
    """Compute the metrics on forecast"""
    metrics = {"MAE": MAE(), "MSE": MSE(), "SMAPE": SMAPE()}
    results = dict()
    for name, metric in metrics.items():
        results[name] = metric(y_true=test, y_pred=forecast)[segment]
    return results

def test_transforms(transforms=[]):
    """Run the experiment on the list of transforms"""
    df = pd.read_csv("data/example_dataset.csv")
    df = df[df["segment"] == segment]
    ts = TSDataset(df, freq="D")
    train, test = ts.train_test_split(
        train_start="2019-05-20",
        train_end="2019-07-10",
        test_start="2019-07-11",
        test_end="2019-08-09",
    )

    model = Pipeline(model=MovingAverageModel(window=30), transforms=transforms, horizon=30)
    model.fit(train)
    forecast = model.forecast()
    metrics = get_metrics(forecast, test)
    return metrics

Results on the uncleaned dataset:

test_transforms()

{'MAE': 40.08799715116714,
 'MSE': 1704.8554888537708,
 'SMAPE': 27.36913416466395}

Results on the cleaned dataset(we impute outliers in train part and predict the test part):

transforms = [outliers_remover, outliers_imputer]
test_transforms(transforms)

{'MAE': 11.606826505006092,
 'MSE': 196.5736131226554,
 'SMAPE': 8.94919204071121}

As you can see, there is significant improvement in the metrics in our case. However, the points that we defined as the outliers may occur to be the points of complex timeseries behavior. In this case, outliers imputation might make the forecast worse.