A snippet on how to use the custom RandomForest class

rf_growth.py

# Data wrangling 
import pandas as pd 

# Accuracy metrics 
from sklearn.metrics import precision_score, recall_score

# Reading data for classification 
d = pd.read_csv("data/random_forest/telecom_churn.csv")

# Setting the features used 
features = [
    'AccountWeeks',
    'DataUsage',
    'DayMins',
    'DayCalls',
    'MonthlyCharge',
    'OverageFee',
    'RoamMins'
]

# Initiating the random forest object 
rf = RandomForestClassifier(
    Y=d['Churn'], 
    X=d[features],
    min_samples_split=5,
    max_depth=4,
    n_trees=10,
    X_features_fraction=0.5
    )

# Growing the random forest 
rf.grow_random_forest()

# Printing out the trees 
rf.print_trees()

# Making predictions
yhat = rf.predict(d[features])
d['yhat'] = yhat 

# Measurring accuracy
print(f"The training precision: {precision_score(d['Churn'], d['yhat'])}")
print(f"The training recall: {recall_score(d['Churn'], d['yhat'])}")