Data Regressors
Ongoing project that performs regression of a continuous variable (oil rate) from a set of associated variables using different ML and DL techniques and approaches, such as linear regression, time series, ensemble methods or neural networks.
Analysis
1. Data Exploration and Profiling
- Data Exploration
- Descriptive Statistics
- Data Profiling
2. Time Series
- Visual Analytics
- Base Line
- Time Series Analysis
- Box-Jenkins Analysis
- Holt-Winters Analysis
- Compare Models
3. Random Forest Regression
- Quick example
4. NN Regression
- Load and show data
- Prepare the data to Learn
- Create Train/Validation/Test datasets
- Train Model
- Make predictions and calculate error
- Plot Results
5. Random Forest Regression
- Quick example
Data
The dataset used is composed of a group a group of operational and PVT variables for oil wells.
| # | Variable | Type | Unit | Data Type |
|---|---|---|---|---|
| 1 | WellID | Numerical | ||
| 2 | Date | Date | ||
| 3 | MethodID | Operational | Categorical | |
| 4 | CHP | Operational | psi | Numerical |
| 5 | THP | Operational | psi | Numerical |
| 6 | Temp | Operational | Fahrenheit | Numerical |
| 7 | Choke | Operational | inches | Numerical |
| 8 | Qinj | Operational | Mscf | Numerical |
| 9 | Bo | PVT | Numerical | |
| 10 | Zed | PVT | Numerical | |
| 11 | SpgO | PVT | Numerical | |
| 12 | SpgGP | PVT | Numerical | |
| 13 | Rel_Oper_Press | Operational | Numerical | |
| 14 | Rel_Crit_Press | Operational | Numerical | |
| 15 | WC | Operational | Numerical | |
| 16 | Test_Oil | Well Test | bbls | Numerical |
All datasets used in this project can be viewed in this folder.
Dependencies
To install these packages with conda, run the following commands:
conda install -c conda-forge pandas-profiling
conda install -c conda-forge keras
conda install -c anaconda pydot
Contributing and Feedback
Any kind of feedback/suggestions would be greatly appreciated (algorithm design, documentation, improvement ideas, spelling mistakes, etc…). If you want to make a contribution to the course you can do it through a PR.
Authors
- Created by Segura Tinoco, Andrés
- Created on Oct 4, 2019
- Updated on Apr 15, 2022
License
This project is licensed under the terms of the MIT license.