Regression¶

---

This example shows how to use ATOM to apply PCA on the data and run a regression pipeline.

Download the abalone dataset from https://archive.ics.uci.edu/ml/datasets/Abalone. The goal of this dataset is to predict the rings (age) of abalone shells from physical measurements.

Load the data¶

In [14]:

# Import packages
import pandas as pd
from atom import ATOMRegressor

# Import packages import pandas as pd from atom import ATOMRegressor

In [15]:

# Load the data
X = pd.read_csv("./datasets/abalone.csv")

# Let's have a look
X.head()

# Load the data X = pd.read_csv("./datasets/abalone.csv") # Let's have a look X.head()

Out[15]:

	Sex	Length	Diameter	Height	Whole weight	Shucked weight	Viscera weight	Shell weight	Rings
0	M	0.455	0.365	0.095	0.5140	0.2245	0.1010	0.150	15
1	M	0.350	0.265	0.090	0.2255	0.0995	0.0485	0.070	7
2	F	0.530	0.420	0.135	0.6770	0.2565	0.1415	0.210	9
3	M	0.440	0.365	0.125	0.5160	0.2155	0.1140	0.155	10
4	I	0.330	0.255	0.080	0.2050	0.0895	0.0395	0.055	7

In [16]:

# Initialize atom for regression tasks
atom = ATOMRegressor(X, "Rings", verbose=2, warnings=False, random_state=42)

# Initialize atom for regression tasks atom = ATOMRegressor(X, "Rings", verbose=2, warnings=False, random_state=42)

<< ================== ATOM ================== >>
Algorithm task: regression.

Dataset stats ==================== >>
Shape: (4177, 9)
Scaled: False
Categorical features: 1 (12.5%)
Outlier values: 192 (0.6%)
-------------------------------------
Train set size: 3342
Test set size: 835

In [17]:

# Encode the categorical features
atom.encode()

# Encode the categorical features atom.encode()

Fitting Encoder...
Encoding categorical columns...
 --> OneHot-encoding feature Sex. Contains 3 classes.

In [18]:

# Plot the dataset's correlation matrix
atom.plot_correlation()

# Plot the dataset's correlation matrix atom.plot_correlation()

In [19]:

# Apply PCA for dimensionality reduction
atom.feature_selection(strategy="pca", n_features=6)

# Apply PCA for dimensionality reduction atom.feature_selection(strategy="pca", n_features=6)

Fitting FeatureSelector...
Performing feature selection ...
 --> Applying Principal Component Analysis...
   >>> Scaling features...
   >>> Explained variance ratio: 0.976

In [20]:

# Note that the fetaures are automatically renamed to Component 1, 2, etc...
atom.columns

# Note that the fetaures are automatically renamed to Component 1, 2, etc... atom.columns

Out[20]:

['component 1',
 'component 2',
 'component 3',
 'component 4',
 'component 5',
 'component 6',
 'Rings']

In [21]:

# Use the plotting methods to see the retained variance ratio
atom.plot_pca()

# Use the plotting methods to see the retained variance ratio atom.plot_pca()

In [22]:

atom.plot_components(figsize=(8, 6))

atom.plot_components(figsize=(8, 6))

Run the pipeline¶

In [23]:

atom.run(
    models=["Tree", "Bag", "ET"],
    metric="MSE",
    n_calls=5,
    n_initial_points=2,
    bo_params={"base_estimator": "GBRT"},
    n_bootstrap=5,
)

atom.run( models=["Tree", "Bag", "ET"], metric="MSE", n_calls=5, n_initial_points=2, bo_params={"base_estimator": "GBRT"}, n_bootstrap=5, )

Training ========================= >>
Models: Tree, Bag, ET
Metric: neg_mean_squared_error


Running BO for Decision Tree...
| call             |   criterion | splitter | max_depth | min_samples_split | min_samples_leaf | max_features | ccp_alpha | neg_mean_squared_error | best_neg_mean_squared_error |    time | total_time |
| ---------------- | ----------- | -------- | --------- | ----------------- | ---------------- | ------------ | --------- | ---------------------- | --------------------------- | ------- | ---------- |
| Initial point 1  | absolute_.. |   random |         7 |                 8 |               19 |         auto |    0.0161 |                 -7.725 |                      -7.725 |  0.109s |     0.125s |
| Initial point 2  | absolute_.. |     best |         6 |                 3 |               12 |         None |       0.0 |                -8.3733 |                      -7.725 |  0.219s |     0.469s |
| Iteration 3      |     poisson |   random |         7 |                 7 |               17 |         auto |    0.0177 |                -8.2151 |                      -7.725 |  0.047s |     0.703s |
| Iteration 4      |     poisson |   random |         6 |                 9 |               19 |         None |    0.0018 |                -9.7263 |                      -7.725 |  0.047s |     0.922s |
| Iteration 5      | friedman_.. |   random |         7 |                10 |               19 |         auto |    0.0093 |                -7.0726 |                     -7.0726 |  0.047s |     1.156s |
Bayesian Optimization ---------------------------
Best call --> Iteration 5
Best parameters --> {'criterion': 'friedman_mse', 'splitter': 'random', 'max_depth': 7, 'min_samples_split': 10, 'min_samples_leaf': 19, 'max_features': 'auto', 'ccp_alpha': 0.0093}
Best evaluation --> neg_mean_squared_error: -7.0726
Time elapsed: 1.344s
Fit ---------------------------------------------
Train evaluation --> neg_mean_squared_error: -6.3272
Test evaluation --> neg_mean_squared_error: -5.4156
Time elapsed: 0.016s
Bootstrap ---------------------------------------
Evaluation --> neg_mean_squared_error: -5.4807 ± 0.1807
Time elapsed: 0.031s
-------------------------------------------------
Total time: 1.391s


Running BO for Bagging...
| call             | n_estimators | max_samples | max_features | bootstrap | bootstrap_features | neg_mean_squared_error | best_neg_mean_squared_error |    time | total_time |
| ---------------- | ------------ | ----------- | ------------ | --------- | ------------------ | ---------------------- | --------------------------- | ------- | ---------- |
| Initial point 1  |          112 |         0.9 |          0.6 |     False |              False |                 -6.007 |                      -6.007 |  1.172s |     1.172s |
| Initial point 2  |          131 |         0.5 |          0.5 |     False |              False |                 -7.047 |                      -6.007 |  0.828s |     2.125s |
| Iteration 3      |           50 |         0.9 |          0.6 |     False |               True |                -5.3357 |                     -5.3357 |  0.484s |     2.828s |
| Iteration 4      |           74 |         0.5 |          0.5 |     False |               True |                -6.1811 |                     -5.3357 |  0.391s |     3.766s |
| Iteration 5      |           18 |         0.8 |          0.6 |      True |              False |                -6.3648 |                     -5.3357 |  0.159s |     4.160s |
Bayesian Optimization ---------------------------
Best call --> Iteration 3
Best parameters --> {'n_estimators': 50, 'max_samples': 0.9, 'max_features': 0.6, 'bootstrap': False, 'bootstrap_features': True}
Best evaluation --> neg_mean_squared_error: -5.3357
Time elapsed: 4.332s
Fit ---------------------------------------------
Train evaluation --> neg_mean_squared_error: -0.0867
Test evaluation --> neg_mean_squared_error: -4.9533
Time elapsed: 0.563s
Bootstrap ---------------------------------------
Evaluation --> neg_mean_squared_error: -5.1436 ± 0.0702
Time elapsed: 2.578s
-------------------------------------------------
Total time: 7.472s


Running BO for Extra-Trees...
| call             | n_estimators |     criterion | max_depth | min_samples_split | min_samples_leaf | max_features | bootstrap | ccp_alpha | max_samples | neg_mean_squared_error | best_neg_mean_squared_error |    time | total_time |
| ---------------- | ------------ | ------------- | --------- | ----------------- | ---------------- | ------------ | --------- | --------- | ----------- | ---------------------- | --------------------------- | ------- | ---------- |
| Initial point 1  |          112 | absolute_er.. |         1 |                 9 |                7 |          0.6 |      True |    0.0161 |         0.6 |               -10.2144 |                    -10.2144 |  0.406s |     0.422s |
| Initial point 2  |          369 | absolute_er.. |      None |                 3 |               12 |         None |      True |    0.0347 |         0.7 |                -9.2503 |                     -9.2503 |  5.063s |     5.563s |
| Iteration 3      |          383 | absolute_er.. |      None |                 6 |               20 |         None |      True |    0.0271 |         0.8 |                -6.4866 |                     -6.4866 |  5.989s |    11.724s |
| Iteration 4      |          412 | absolute_er.. |         1 |                13 |               17 |          0.6 |      True |    0.0282 |         0.8 |               -10.0739 |                     -6.4866 |  1.866s |    13.862s |
| Iteration 5      |          125 | squared_error |      None |                14 |               20 |         None |      True |    0.0156 |         0.9 |                -6.2858 |                     -6.2858 |  0.220s |    14.291s |
Bayesian Optimization ---------------------------
Best call --> Iteration 5
Best parameters --> {'n_estimators': 125, 'criterion': 'squared_error', 'max_depth': None, 'min_samples_split': 14, 'min_samples_leaf': 20, 'max_features': None, 'bootstrap': True, 'ccp_alpha': 0.0156, 'max_samples': 0.9}
Best evaluation --> neg_mean_squared_error: -6.2858
Time elapsed: 14.482s
Fit ---------------------------------------------
Train evaluation --> neg_mean_squared_error: -6.0139
Test evaluation --> neg_mean_squared_error: -4.9885
Time elapsed: 0.206s
Bootstrap ---------------------------------------
Evaluation --> neg_mean_squared_error: -4.984 ± 0.0229
Time elapsed: 0.998s
-------------------------------------------------
Total time: 15.686s


Final results ==================== >>
Duration: 24.549s
-------------------------------------
Decision Tree --> neg_mean_squared_error: -5.4807 ± 0.1807 ~
Bagging       --> neg_mean_squared_error: -5.1436 ± 0.0702 ~
Extra-Trees   --> neg_mean_squared_error: -4.984 ± 0.0229 ~ !

Analyze the results¶

In [24]:

# Use the errors or residuals plots to check the model performances
atom.plot_residuals()

# Use the errors or residuals plots to check the model performances atom.plot_residuals()

In [25]:

atom.plot_errors()

atom.plot_errors()

In [26]:

# Analyze the relation between the target response and the features
atom.n_jobs = 8  # The method can be slow...
atom.ET.plot_partial_dependence(columns=(0, (2, 3)), figsize=(12, 5))

# Analyze the relation between the target response and the features atom.n_jobs = 8 # The method can be slow... atom.ET.plot_partial_dependence(columns=(0, (2, 3)), figsize=(12, 5))