/N_step_oil_rate_prediction

Here is an example of a solution to oil rate prediction on N_steps

Primary LanguageJupyter Notebook

Oil rate prediction on N steps

Here is an example of a solution to oil rate prediction on 90 days ahead applied to 106 wells Solution is based on Catboost regression model with deep feature engineering research. Features are aimed to analyze sliding window statistics

Validation Catboost model RMSE = 2.78

Validation constant prediction RMSE = 3.2

Predictions

Well 28

image

Well 11

image