[ENH] Functionality to collate and summarise classifier parameters over datasets

[TonyBagnall]

Describe the feature or idea you want to propose

Line two in the results files contains a dictionary for classifier parameters

e.g.
{'batch_size': 100, 'contract_max_n_shapelet_samples': inf, 'estimator': None, 'max_shapelet_length': None, 'max_shapelets': 1000, 'n_jobs': 1, 'n_shapelet_samples': 50000, 'random_state': 1, 'save_transformed_data': False, 'time_limit_in_minutes': 0, 'transform_limit_in_minutes': 0}

it would be good to firstly list variables set in fit here, and secondly be able to summarise this over resamples and problems to help understand how they are operating

Describe your proposed solution

so input a dictionary, generate a file for each parameter over resamples

for example collate_paras("estimator",["para1", "para2"]) create files para1 with 112x30 of para1 and para2

Other comments and alternatives considered

No response