Models

pyfbad.models.models.IsolationForestModel

That model includes classic Isolation Forest (IF) Model. It has train_model() method

>>> train_model(self, df_model, contamination_value=float(0.06))

It has default contamination value (float(0.06)) and it trains the model and make prediction with given dataframe.

That model includes classic Isolation Forest (IF) Model. It has train_model() method

>>> train_model(self, df_model, contamination_value=float(0.06))

It has default contamination value (float(0.06)) and it trains the model and make prediction with given dataframe.

That model developed by facebook and that is our first model we implemented on pyfbad.

>>> train_model(self, df_model)

Train a Prophet model with given dataframe

>>> train_forecast(self, forecast, bound_coefficient)

Tries to predict anomalies based on training results. bound_coefficient (float): optimization coefficient for anomaly number

>>> get_anomalies(self, model_result, anomaly_number_level="Low")

Tries to predict anomalies based on number level for each coefficients. anomaly_number_level (str): detected total anomaly number, high or low

>>> find_optimum_anomalies(self, anomaly_table, results):

Tries to find best coefficient for getting optimum anomalies.

>>> train_model(self, df_model, cluster_number, random_state=7)

Train a Gaussian Mixture model with given dataframe.

>>> get_all_models(self, data)

Train a Gaussian Mixture model with different cluster number values.

>>> find_best_model(self, bic_table, models)

Find best gmm model wit respect to different cluster numbers.

>>> train_forecast(self, gmm_model, model_data, anomaly_percent)

Forecasting anomalies using found best gmm model. anomaly_percent (int): threshold value for number of detected anomalies