AutomaticTimeSeriesFitAndSave

class hana_ai.tools.df_tools.automatic_timeseries_tools.AutomaticTimeSeriesFitAndSave(connection_context: ConnectionContext, return_direct: bool = False)

This tool fits a time series model and saves it in the model storage.

Parameters:

connection_contextConnectionContext

Connection context to the HANA database.

Returns:

str

The JSON string of the trained table, model storage name, and model storage version.

Note

args_schema is used to define the schema of the inputs as follows:

Field	Description
fit_select_statement	The SQL select statement of the input dataset to fit the model. If not provided, ask the user. Do not guess.
name	The name of the model in model storage. If not provided, ask the user. Do not guess.
version	The version of the model in model storage, it is optional
scorings	The scorings for the model, e.g. {'MAE':-1.0, 'EVAR':1.0} and it supports EVAR, MAE, MAPE, MAX_ERROR, MSE, R2, RMSE, WMAPE, LAYERS, SPEC, TIME, and it is optional
generations	The number of iterations of the pipeline optimization., it is optional
population_size	The number of individuals in the population., it is optional
offspring_size	The number of children to produce at each generation., it is optional
elite_number	The number of the best individuals to select for the next generation., it is optional
min_layer	The minimum number of layers in the pipeline., it is optional
max_layer	The maximum number of layers in the pipeline., it is optional
mutation_rate	The mutation rate., it is optional
crossover_rate	The crossover rate., it is optional
random_seed	The random seed., it is optional
config_dict	The configuration dictionary for the searching space, it is optional
progress_indicator_id	The progress indicator id, it is optional
fold_num	The number of folds for cross validation, it is optional
resampling_method	The resampling method for cross validation from {'rocv', 'block'}, it is optional
max_eval_time_mins	The maximum evaluation time in minutes, it is optional
early_stop	Stop optimization progress when the best pipeline is not updated for the give consecutive generations and 0 means there is no early stop, and it is optional
percentage	The percentage of the data to be used for training, it is optional
gap_num	The number of samples to exclude from the end of each train set before the test set, it is optional
connections	The connections for the model, it is optional
alpha	The rejection probability in connection optimization, it is optional
delta	The minimum improvement in connection optimization, it is optional
top_k_connections	The number of top connections to keep in connection optimization, it is optional
top_k_pipelines	The number of top pipelines to keep in pipeline optimization, it is optional
fine_tune_pipline	Whether to fine tune the pipeline, it is optional
fine_tune_resource	The resource for fine tuning, it is optional
key	The key of the dataset. If not provided, ask the user. Do not guess.
endog	The endog of the dataset. If not provided, ask the user. Do not guess.
exog	The exog of the dataset, it is optional
categorical_variable	The categorical variable of the dataset, it is optional
background_size	The amount of background data in Kernel SHAP. Its value should not exceed the number of rows in the training data, it is optional
background_sampling_seed	The seed for sampling the background data in Kernel SHAP, it is optional
use_explain	Whether to use explain, it is optional
workload_class	The workload class for fitting the model, it is optional

name: str

Name of the tool.

description: str

Description of the tool.

connection_context: ConnectionContext

Connection context to the HANA database.