Supervised metric and not metric space learning methods for multiple purposes in Machine Learning
- supervised similarity matching
- nomparametric density estimation
- ensemble learning
pip install your_project_name
import seaborn as sns
from sklearn.datasets import fetch_california_housing
from heartwood.kernel import JaccardForestKernel, MLPKernel, DiscretizedTargetKernel, make_bimodal_assymetric_regressionX_bimodal, y_bimodal = make_bimodal_assymetric_regression(10_000)housing = fetch_california_housing(as_frame = True)['frame']numerical_cols = housing.columns[:-1]
target_col = housing.columns[-1]housing
<style scoped>
.dataframe tbody tr th:only-of-type {
vertical-align: middle;
}
</style>
.dataframe tbody tr th {
vertical-align: top;
}
.dataframe thead th {
text-align: right;
}
| MedInc | HouseAge | AveRooms | AveBedrms | Population | AveOccup | Latitude | Longitude | MedHouseVal | |
|---|---|---|---|---|---|---|---|---|---|
| 0 | 8.3252 | 41.0 | 6.984127 | 1.023810 | 322.0 | 2.555556 | 37.88 | -122.23 | 4.526 |
| 1 | 8.3014 | 21.0 | 6.238137 | 0.971880 | 2401.0 | 2.109842 | 37.86 | -122.22 | 3.585 |
| 2 | 7.2574 | 52.0 | 8.288136 | 1.073446 | 496.0 | 2.802260 | 37.85 | -122.24 | 3.521 |
| 3 | 5.6431 | 52.0 | 5.817352 | 1.073059 | 558.0 | 2.547945 | 37.85 | -122.25 | 3.413 |
| 4 | 3.8462 | 52.0 | 6.281853 | 1.081081 | 565.0 | 2.181467 | 37.85 | -122.25 | 3.422 |
| ... | ... | ... | ... | ... | ... | ... | ... | ... | ... |
| 20635 | 1.5603 | 25.0 | 5.045455 | 1.133333 | 845.0 | 2.560606 | 39.48 | -121.09 | 0.781 |
| 20636 | 2.5568 | 18.0 | 6.114035 | 1.315789 | 356.0 | 3.122807 | 39.49 | -121.21 | 0.771 |
| 20637 | 1.7000 | 17.0 | 5.205543 | 1.120092 | 1007.0 | 2.325635 | 39.43 | -121.22 | 0.923 |
| 20638 | 1.8672 | 18.0 | 5.329513 | 1.171920 | 741.0 | 2.123209 | 39.43 | -121.32 | 0.847 |
| 20639 | 2.3886 | 16.0 | 5.254717 | 1.162264 | 1387.0 | 2.616981 | 39.37 | -121.24 | 0.894 |
20640 rows × 9 columns
sns.distplot(housing.iloc[:,-1])<AxesSubplot:xlabel='MedHouseVal', ylabel='Density'>
uses forest node embeddings and jaccard similarity to sample points
from sklearn.ensemble import ExtraTreesRegressor, RandomForestRegressor
from sklearn.model_selection import train_test_split
learner = JaccardForestKernel(
RandomForestRegressor(n_jobs = -1),
fit_neighbors_index = True,
verbose = True
)housing_train, housing_test = train_test_split(housing, train_size = 0.8)learner.fit(
housing_train[numerical_cols],
housing_train[target_col],
save_values = housing_train
)JaccardForestKernel(estimator=RandomForestRegressor(n_jobs=-1), verbose=True)
queries = learner.query(housing_test[numerical_cols])kNN time total=1.645799 (sec), per query=0.000399 (sec), per query adjusted for thread number=0.001595 (sec)
import numpy as np
import matplotlib.pyplot as plt
idx = np.random.choice(len(housing_test))
q = queries[0][idx]
q['similarity'] = queries[1][idx]
sns.distplot(q[target_col])
plt.axvline(housing_test.iloc[idx][target_col], color = 'r')<matplotlib.lines.Line2D at 0x17834339e08>
q[:10]
<style scoped>
.dataframe tbody tr th:only-of-type {
vertical-align: middle;
}
</style>
.dataframe tbody tr th {
vertical-align: top;
}
.dataframe thead th {
text-align: right;
}
| MedInc | HouseAge | AveRooms | AveBedrms | Population | AveOccup | Latitude | Longitude | MedHouseVal | similarity | |
|---|---|---|---|---|---|---|---|---|---|---|
| 2552 | 2.2227 | 30.0 | 5.278552 | 1.019499 | 990.0 | 2.757660 | 40.77 | -124.17 | 0.813 | 0.907104 |
| 2648 | 2.2357 | 30.0 | 5.152406 | 1.050802 | 996.0 | 2.663102 | 40.50 | -124.10 | 0.723 | 0.913043 |
| 2636 | 2.0938 | 39.0 | 5.594340 | 1.122642 | 539.0 | 2.542453 | 40.59 | -124.15 | 0.796 | 0.924731 |
| 2575 | 2.2478 | 31.0 | 5.123810 | 1.100000 | 1259.0 | 2.997619 | 40.80 | -124.13 | 0.811 | 0.924731 |
| 19788 | 2.4792 | 26.0 | 6.129121 | 1.156593 | 987.0 | 2.711538 | 40.71 | -122.95 | 0.888 | 0.941799 |
| 2650 | 2.3947 | 38.0 | 5.536160 | 1.062344 | 1041.0 | 2.596010 | 40.44 | -124.09 | 0.705 | 0.947368 |
| 19763 | 2.2500 | 30.0 | 5.861862 | 1.192192 | 961.0 | 2.885886 | 40.02 | -122.18 | 0.682 | 0.963731 |
| 2662 | 2.0870 | 33.0 | 6.141975 | 1.216049 | 1221.0 | 2.512346 | 40.12 | -123.82 | 0.824 | 0.969072 |
| 18794 | 2.1250 | 22.0 | 6.856688 | 1.468153 | 903.0 | 2.875796 | 40.40 | -122.76 | 1.232 | 0.969072 |
| 1860 | 2.1336 | 19.0 | 5.893993 | 1.144876 | 841.0 | 2.971731 | 41.92 | -124.16 | 0.750 | 0.974359 |
conditional target distribution could be bimodal, making regular regression faill. discretizing and running a classification might help
from sklearn.linear_model import LogisticRegression
from sklearn.pipeline import make_pipeline
from sklearn.preprocessing import RobustScalerdisc_learner = DiscretizedTargetKernel(
estimator = make_pipeline(RobustScaler(), LogisticRegression()),
n_bins = 15,
strategy = 'uniform'
)disc_learner.fit(
housing_train[numerical_cols],
housing_train[[target_col]],
save_values = housing_train
)DiscretizedTargetKernel(estimator=Pipeline(steps=[('robustscaler',
RobustScaler()),
('logisticregression',
LogisticRegression())]),
n_bins=15, strategy='uniform')
disc_queries = disc_learner.query(housing_test[numerical_cols])idx = np.random.choice(len(housing_test))
q = disc_queries[0][idx]
q['kldivergence'] = disc_queries[1][idx]
sns.distplot(q[target_col])
plt.axvline(housing_test.iloc[idx][target_col], color = 'r')<matplotlib.lines.Line2D at 0x1782f883fc8>
q[:10]
<style scoped>
.dataframe tbody tr th:only-of-type {
vertical-align: middle;
}
</style>
.dataframe tbody tr th {
vertical-align: top;
}
.dataframe thead th {
text-align: right;
}
| MedInc | HouseAge | AveRooms | AveBedrms | Population | AveOccup | Latitude | Longitude | MedHouseVal | kldivergence | |
|---|---|---|---|---|---|---|---|---|---|---|
| 12947 | 4.8661 | 24.0 | 6.358456 | 1.053309 | 1336.0 | 2.455882 | 38.64 | -121.28 | 1.862 | 0.002280 |
| 19666 | 4.6010 | 34.0 | 6.320329 | 0.979466 | 1226.0 | 2.517454 | 37.51 | -120.83 | 1.500 | 0.002638 |
| 16450 | 4.1190 | 38.0 | 5.506876 | 0.982318 | 1223.0 | 2.402750 | 38.14 | -121.28 | 1.288 | 0.003778 |
| 12731 | 5.0508 | 32.0 | 7.575688 | 1.100917 | 1185.0 | 2.717890 | 38.59 | -121.36 | 2.257 | 0.003865 |
| 20559 | 4.7222 | 30.0 | 6.424691 | 0.990123 | 981.0 | 2.422222 | 38.67 | -121.79 | 1.672 | 0.004230 |
| 12727 | 4.4643 | 35.0 | 5.710059 | 0.970414 | 805.0 | 2.381657 | 38.60 | -121.36 | 1.330 | 0.004297 |
| 1534 | 4.2625 | 37.0 | 5.925795 | 0.950530 | 689.0 | 2.434629 | 37.89 | -122.05 | 3.106 | 0.004309 |
| 19057 | 4.2813 | 33.0 | 6.340426 | 1.117021 | 212.0 | 2.255319 | 38.13 | -121.84 | 0.813 | 0.004316 |
| 18894 | 4.0938 | 37.0 | 5.979592 | 1.046382 | 1325.0 | 2.458256 | 38.13 | -122.24 | 1.269 | 0.005845 |
| 2040 | 4.9196 | 27.0 | 6.184783 | 0.906522 | 1219.0 | 2.650000 | 36.73 | -119.67 | 1.179 | 0.006963 |