Hyperameter_tuning
Hyperameter Tunning
- We settled with the Catboost model for good Quality prediction and low latency prediction
This is the table of hyperparameters and their performance metric
| ('number', '') | ('state', '') | ('value', '') | ('datetime_start', '') | ('datetime_complete', '') | ('params', 'bagging_temperature') | ('params', 'boosting_type') | ('params', 'bootstrap_type') | ('params', 'depth') | ('params', 'learning_rate') | ('params', 'objective') | ('params', 'subsample') | ('system_attrs', '_number') | ('system_attrs', 'fail_reason') | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 0 | TrialState.FAIL | nan | 2020-01-02 20:21:11.863237 | 2020-01-02 20:21:12.418973 | nan | nan | nan | 6 | nan | RMSE | nan | 0 | Setting status of trial#0 as TrialState.FAIL because of the following error: TypeError('suggest_uniform() takes 4 positional arguments but 8 were given',) |
| 1 | 1 | TrialState.FAIL | nan | 2020-01-02 20:21:38.181381 | 2020-01-02 20:21:38.818264 | nan | nan | nan | 7 | nan | MAE | nan | 1 | Setting status of trial#1 as TrialState.FAIL because of the following error: TypeError('suggest_uniform() takes 4 positional arguments but 6 were given',) |
| 2 | 2 | TrialState.FAIL | nan | 2020-01-02 20:22:14.453568 | 2020-01-02 20:22:15.072724 | nan | nan | nan | 2 | nan | RMSE | nan | 2 | Setting status of trial#2 as TrialState.FAIL because of the following error: TypeError('suggest_uniform() takes 4 positional arguments but 5 were given',) |
| 3 | 3 | TrialState.COMPLETE | 42.7238 | 2020-01-02 20:22:23.798479 | 2020-01-02 20:22:39.130715 | nan | Plain | MVS | 11 | 0.489043 | MAE | nan | 3 | nan |
| 4 | 4 | TrialState.COMPLETE | 50.5916 | 2020-01-02 20:22:39.215907 | 2020-01-02 20:22:42.882038 | nan | Plain | Bernoulli | 5 | 0.529281 | RMSE | 0.607423 | 4 | nan |
| 5 | 5 | TrialState.COMPLETE | 42.5857 | 2020-01-02 20:22:42.963390 | 2020-01-02 20:23:43.219132 | nan | Plain | Bernoulli | 3 | 0.683597 | MAE | 0.374904 | 5 | nan |
| 6 | 6 | TrialState.COMPLETE | 46.9529 | 2020-01-02 20:23:43.311247 | 2020-01-02 20:24:41.338719 | nan | Plain | MVS | 1 | 0.0137888 | MAE | nan | 6 | nan |
| 7 | 7 | TrialState.COMPLETE | 44.549 | 2020-01-02 20:24:41.431578 | 2020-01-02 20:25:29.054821 | 1.91368 | Ordered | Bayesian | 1 | 0.384213 | MAE | nan | 7 | nan |
| 8 | 8 | TrialState.COMPLETE | 41.917 | 2020-01-02 20:25:29.147049 | 2020-01-02 20:26:22.714733 | nan | Ordered | MVS | 5 | 0.226351 | MAE | nan | 8 | nan |
| 9 | 9 | TrialState.COMPLETE | 50.1082 | 2020-01-02 20:26:22.825344 | 2020-01-02 20:26:34.244275 | nan | Plain | MVS | 12 | 0.547259 | RMSE | nan | 9 | nan |
| 10 | 10 | TrialState.COMPLETE | 52.9363 | 2020-01-02 20:26:34.396108 | 2020-01-02 20:26:37.911406 | nan | Plain | Bernoulli | 5 | 0.603843 | RMSE | 0.303779 | 10 | nan |
| 11 | 11 | TrialState.COMPLETE | 51.0596 | 2020-01-02 20:26:38.002327 | 2020-01-02 20:26:46.647710 | 8.14127 | Ordered | Bayesian | 4 | 0.754896 | RMSE | nan | 11 | nan |
| 12 | 12 | TrialState.COMPLETE | 50.528 | 2020-01-02 20:26:46.733670 | 2020-01-02 20:26:51.632994 | nan | Ordered | Bernoulli | 4 | 0.51049 | RMSE | 0.546119 | 12 | nan |
| 13 | 13 | TrialState.COMPLETE | 42.2105 | 2020-01-02 20:26:51.725349 | 2020-01-02 20:27:16.891392 | nan | Ordered | MVS | 9 | 0.212695 | MAE | nan | 13 | nan |
| 14 | 14 | TrialState.COMPLETE | 43.212 | 2020-01-02 20:27:16.980661 | 2020-01-02 20:27:50.826629 | nan | Ordered | MVS | 9 | 0.209423 | MAE | nan | 14 | nan |
| 15 | 15 | TrialState.COMPLETE | 42.605 | 2020-01-02 20:27:50.911060 | 2020-01-02 20:28:22.338762 | nan | Ordered | MVS | 8 | 0.171057 | MAE | nan | 15 | nan |
| 16 | 16 | TrialState.COMPLETE | 42.9026 | 2020-01-02 20:28:22.445095 | 2020-01-02 20:28:46.446403 | nan | Ordered | MVS | 8 | 0.29398 | MAE | nan | 16 | nan |
| 17 | 17 | TrialState.COMPLETE | 42.9598 | 2020-01-02 20:28:46.558330 | 2020-01-02 20:30:10.208094 | nan | Ordered | MVS | 10 | 0.0427293 | MAE | nan | 17 | nan |