I'm currently learning Scikit-learn (please don't scold me), and I'm a bit confused about the process regarding ColumnTransformer, training and prediction. I have a data set with features such as Gender, Married, Graduation status, Loan Amount, Income, etc. The data set has a mix of objects (strings) and integer values, but I would say the majority is objects. From my understanding, I need to convert the objects to integer values before training a model, and I do so using ColumnTransformer. But the process of training the model makes me a bit confused. This is my current code:
import pandas as pd
import numpy as np
from sklearn.tree import DecisionTreeClassifier
from sklearn.model_selection import cross_val_score, train_test_split
from sklearn.metrics import accuracy_score
from sklearn.preprocessing import OneHotEncoder
from sklearn.compose import make_column_transformer
df = pd.read_csv("loan_data.csv", sep=",")
df.replace("", np.nan, inplace=True)
df.dropna(inplace=True)
df = df.drop(columns=["Loan_ID"])
X = df.drop(columns=["LoanAmount"])
y = df["LoanAmount"]
loan_categories = ["Gender", "Married", "Dependents", "Education", "Self_Employed", "Property_Area", "Loan_Status"]
ohe = OneHotEncoder()
ct = make_column_transformer (
(ohe, loan_categories),
remainder="passthrough")
ct.fit_transform(X)
And then comes my confusion with train_test_split. Am I supposed to train_test_split before passing X to fit_transform, or would this happen now after I have defined ct?
The rest of my code would look something like this:
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.2, random_state = 42)
model = DecisionTreeClassifier()
model.fit(X_train, y_train)
predictions = model.predict(X_test)
score = accuracy_score(y_test, predictions)