The purpose of this notebook was to develop an automated function to predict the price of a diamond based on its given features (cut, color, dimensions, etc.). We created a machine learning model which can estimate these values. We needed to find continuous data, so we performed a regression task with supervised learning to find the prices. The task was completed with CRISP-DM approach.
To finalize the regression problem, we implemented the following models and achieved these scores:
- polynomial regressor,
- k-nearest neighbors regressor,
- random forest ensemble,
- AdaBoost (adaptive boosting) ensemble.
| Regressor | RMSE | Accuracy | MAE |
|---|---|---|---|
| Polynomial | 978.5624 | 77.6618 | 592.9765 |
| kNN | 646.1506 | 84.0253 | 318.3850 |
| Random Forest | 547.1772 | 85.7200 | 278.1653 |
| AdaBoost | 551.6524 | 84.2118 | 288.2251 |
For the final regressor, we chosen the random forest ensemble and deployed it successfully into a serialized file. The estimating function works well and returns predicted values.