Valerii Pokrovskii

Intro refactoring

2

https://github.com/nccr-itmo/FEDOT/issues/878 issue. Advices are welcome

Due to new changes in core we should propose changes for boosting mutations. Related issues: - https://github.com/nccr-itmo/FEDOT/issues/833 - https://github.com/nccr-itmo/FEDOT/issues/851 What should be done: - [ ] implement boosting mutations as...

Make additional field for decompose models

2

After last changes boosting mutations don't work again due to regression models are not available for classification tasks. The main idea to create new entities for decompose models that will...

Draft of experiment (may be extended) 1) Select 5 datasets for classification/regression/time-series-forecasting 2) For each dataset run such experiment 2.1) Run FEDOT for hour with set of parameters: [0% time...

fix pca

6

This is a 🐛 bug fix. Fixes [#4567](https://github.com/aimclub/FEDOT/issues/1266)

Problem with catboost tuning

5

 example api_classification

Documentation should contains page devoted to metrics. It should consist of * list of metrics for optimization * code shows how to pass custom metric for the API

Forecast length overwriting in Fedot API

3

If we would provide InputData with parameter forecast_length in task_params and wouldn't provide the same task to API, Fedot will faшl with exception Metric can not be evaluated because of:...

Datetime features handling

3

DateTime features like day, month, year, etc can be used as features for time series. 

When forecast_length=1 model.plot_prediction() doesn't show predict  data [df_daily.xlsx](https://github.com/aimclub/FEDOT/files/12802048/df_daily.xlsx) notebook to reproduce https://colab.research.google.com/drive/1iKOufLVJYYf7xTJt0yDklqB7iiTMP6AA?usp=sharing PS need to check is it common for univariate case too