Andrey Stebenkov
Andrey Stebenkov
Solving one of problems from Issue #790. The bug was merging features and targets data. I don't know why, but after the updates release, target array became a 1-dim. Solution...
Solving Issue #826 - Experiment to determine what is better: composing or tuning? Adding some params required to experiments. Changes will not merged in master.
In case data without index column, function `from_csv` set the first column as index and this cause troubles in model fitting. One of the possible solution is to set the...
I tried to make experiment in Fedot for the task of time series forecasting with [Sea level dataset](https://github.com/ITMO-NSS-team/fedot-examples/blob/main/notebooks/data/ts_sea_level.csv). There is ValueError in fit() if you want use np.ndarray array as...
Noticed that the performance of regression models running through PipelineBuilder has decreased compared to the previous version. It is required to compare the last version of framework (0.7.1 or 0.7.2)...
Found an error in case [credit_scroing_problem.py](https://github.com/aimclub/FEDOT/blob/master/cases/credit_scoring/credit_scoring_problem.py). The dataset has numerical features, which each name starts from 'Number'. After [preprocessing](https://github.com/aimclub/FEDOT/blob/c1d2f33a1558c499c494069bd8006445f8b519de/fedot/core/pipelines/pipeline.py#L139) they are transform into categorical and the shape of dataset changes...
~~Adding an ensemble (stacking and bagging) operations to FEDOT~~ UPD: The main task was reformulated and divided into small tasks. In this PR is solving one from this tasks, more...
Adding new features in [Data](https://github.com/aimclub/FEDOT/blob/b937088f583990e7059ce6692187182adc4b219c/fedot/core/data/data.py#L37) class: - [x] Storaging `features_names` in new field - [x] Storaging `features_types` in new field - [x] Storaging `categorical_features` in separetly new field or getting...
After you created pipeline with setting specified 'n_jobs' values in params and trying to fit it, next method [`self.replace_n_jobs_in_nodes(n_jobs)`](https://github.com/aimclub/FEDOT/blob/f02b6262adfbd1d7052e4456b953f2848f6cda42/fedot/core/pipelines/pipeline.py#L185) in `pipeline.fit()` always change `n_jobs` to 1. `self.replace_n_jobs_in_nodes(n_jobs)` requires to avoid...
I would like to be able to get the cleaned data after preprocessing using some simple interface like [fit_transform()](https://auto.gluon.ai/dev/api/autogluon.features.html#autogluon.features.generators.AbstractFeatureGenerator)