AutoML.py

# -*- coding: utf-8 -*-
"""Copy of AutoML.ipynb

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/drive/1n3DBE2sxY28sYUksQH6tofMmqOYF5UEq
"""

!pip install tpot mljar-supervised

from sklearn.metrics import accuracy_score
from sklearn.model_selection import train_test_split
from supervised.automl import AutoML

"""# Options Available

- mode — the package ships with four built-in models. 
  - The Explain mode is ideal for explaining and understanding the data. It results in visualizations of feature importance as well as tree visualizations.
  - The Perform is used when building ML models for production. 
  - The Compete is meant to build models used in machine learning competitions. 
  - The Optuna mode is used to search for highly-tuned ML models.
- algorithms — specifies the algorithms you would like to use. They are usually passed in as a list.
- results_path — the path where the results will be stored
- total_time_limit — the total time in seconds for training the model
- train_ensemble — dictates if an ensemble will be created at the end of the training process
- stack_models — determines if a models stack will be created
- eval_metric — the metric that will be optimized. If auto the logloss is used for classification problems while the rmse is used for regression problems

"""

#automl = AutoML(
    # mode="Explain"
    # algorithms=""
    # results_path="AutoML_22",
    # total_time_limit=30 * 60,
    # train_ensemble=True,
    # stack_models="",
    # eval_metric=""
#)

"""# Healthcare Dataset - [Modified SPARCS dataset](https://raw.githubusercontent.com/hantswilliams/HHA-507-2022/main/autoML/datasets/data_sparcs.csv)

## Load In Dataset
"""

import pandas as pd
sparcs = pd.read_csv('https://raw.githubusercontent.com/hantswilliams/HHA-507-2022/main/autoML/datasets/data_sparcs.csv')
sparcs

sparcs.columns

"""## Potential Variables Of Interest
- Type of Admission (categorical)
- Total Charges (continuous)
- Gender (categorical)
- Race (categorical)
"""

sparcs['Total Charges'].describe()

sparcs['Gender'].describe()

sparcs['Type of Admission'].value_counts()

"""## Create Some Simplified Binary Versions"""

sparcs['Race'] = pd.to_numeric(sparcs['Race'], errors='coerce')
sparcs['sparcs_race'] = sparcs['Race'].apply(lambda x: 'long' if x > 3 else 'short')
sparcs.drop('Race', axis=1, inplace=True)
sparcs['sparcs_race'].value_counts()

"""# MLJar Examples
## Binary Classifier Example 1 - SPARCS
### **Create new model**
"""

x = sparcs.drop(columns=['Race'])

y = sparcs["sparcs_race"]

x

y

x_train, x_test, y_train, y_test = train_test_split(x, y, stratify=y, test_size=0.25)

x_test

automl = AutoML(results_path="Race", mode="Explain")

automl.fit(X_train, y_train)

pred = automl.predict(x_test)
pred

automl.report()

"""### **Test New (not really) Data**"""

# load in the data model 

automl_sparcs_race = AutoML(results_path="Race")

# create a new dataset that follows the same data structure as the training set
X_withrace = sparcs.sample(25)
X_withoutrace = X_withrace.drop(columns=['Race'])

X_withrace

X_withoutrace

predict = automl.predict(X_withoutrace)
predict

# actual values from X_withrace
values_actual = X_withrace['Race'].values.tolist()
values_predicted = predict.tolist()
output = pd.DataFrame({'actual': values_actual, 'predicted': values_predicted})
output

"""## **Binary Classifier Example 2 - GENERIC**"""

import pandas as pd
from supervised.automl import AutoML
import os

from sklearn.metrics import accuracy_score
from sklearn.model_selection import train_test_split

df = pd.read_csv("https://raw.githubusercontent.com/hantswilliams/HHA-507-2022/main/autoML/datasets/data_binary_bank.csv")

X = df[df.columns[:-1]]
y = df["y"]

X

y

X_train, X_test, y_train, y_test = train_test_split(X, y, stratify=y, test_size=0.25)

automl = AutoML(
    # results_path="AutoML_22",
    # total_time_limit=30 * 60,
    # start_random_models=10,
    # hill_climbing_steps=3,
    # top_models_to_improve=3,
    # train_ensemble=True,
    mode="Explain"
)

automl.fit(X_train, y_train)

pred = automl.predict(X_test)
pred
# print("Test accuracy", accuracy_score(y_test, pred["label"]))

automl.report()

"""## **Regression - Example - GENERIC**"""

import numpy as np
import pandas as pd
from supervised.automl import AutoML

df = pd.read_csv("https://raw.githubusercontent.com/hantswilliams/HHA-507-2022/main/autoML/datasets/data_sparcs.csv")
x_cols = [c for c in df.columns if c != "Total Charges"]
X = df[x_cols]
y = df["Total Charges"]

df

x_cols

X

y

automl = AutoML(results_path="sparcs_regression", mode="Explain")
automl.fit(X, y)

df["Predictions"] = automl.predict(X)

print("Predictions")
print(df[["Total Charges", "Predictions"]].head())

"""## **Multiclass Classifier - GENERIC**"""

import pandas as pd
import numpy as np
from supervised.automl import AutoML
import supervised


import warnings

from sklearn import datasets
from sklearn.pipeline import make_pipeline
from sklearn.decomposition import PCA

from supervised import AutoML
from supervised.exceptions import AutoMLException

# warnings.filterwarnings('error')
warnings.filterwarnings(
    "error", category=pd.core.common.SettingWithCopyWarning
)  # message="*ndarray*")

df = pd.read_csv("https://raw.githubusercontent.com/hantswilliams/HHA-507-2022/main/autoML/datasets/data_classes_iris.csv")
X = df[["feature_1", "feature_2", "feature_3", "feature_4"]]
y = df["class"]

df

X

y.value_counts()

automl = AutoML()

automl.fit(X, y)

predictions = automl.predict_all(X)

print(predictions.head())
print(predictions.tail())

print(X.shape)
print(predictions.shape)

"""## **Download Outputs**

### **Binary Classifications Download**
"""

# get current working directory
import os
os.getcwd()

folders = os.listdir()
foldersML = [x for x in folders if x.startswith('Race')]
print(foldersML)

!zip -r /content/Race.zip /content/Race

"""##**Regression Download**"""

os.getcwd()

folders = os.listdir()
foldersML = [x for x in folders if x.startswith('sparcs_regression')]
print(foldersML)

!zip -r /content/Regression.zip /content/sparcs_regression