Medical Cost Prediction

About Dataset Context Machine Learning with R by Brett Lantz is a book that provides an introduction to machine learning using R. As far as I can tell, Packt Publishing does not make its datasets available online unless you buy the book and create a user account which can be a problem if you are checking the book out from the library or borrowing the book from a friend. All of these datasets are in the public domain but simply needed some cleaning up and recoding to match the format in the book.

Content Columns

age: age of primary beneficiary

sex: insurance contractor gender, female, male

bmi: Body mass index, providing an understanding of body, weights that are relatively high or low relative to height, objective index of body weight (kg / m ^ 2) using the ratio of height to weight, ideally 18.5 to 24.9

children: Number of children covered by health insurance / Number of dependents

smoker: Smoking

region: the beneficiary's residential area in the US, northeast, southeast, southwest, northwest.

charges: Individual medical costs billed by health insurance

import numpy as np import pandas as pd import matplotlib.pyplot as plt import seaborn as sns from sklearn.model_selection import train_test_split from sklearn.linear_model import LinearRegression from sklearn import metrics

df=pd.read_csv('./insurance.csv') df.head()

df.shape

df.info( )

df.describe()

#changing categorical varibables to numerical df['sex']=df['sex'].map({'male':1,'female':0}) df['smoker']=df['smoker'].map({'yes':1,'no':0}) df['region']=df['region'].map({'southwest':0,'southeast':1,'northwest':2,'northeast':3})

df.head()

#Age distribution sns.histplot(df.age,bins=10,kde=False,color='g') plt.title("age distribution") plt.xlabel('Age') plt.ylabel('count') plt.show()

#Gender Plot sns.countplot(x='sex',data=df, hue=df.sex) plt.title('gender distribution')

#bmi distribution sns.histplot(df.bmi,bins=20, kde=True,color='red') plt.title('BMI Distribution') plt.xlabel('BMI') plt.ylabel('Count') plt.show()

#child count distribution sns.countplot(x = 'children', data = df,hue=df.children) plt.title('Children Distribution') plt.xlabel('Children') plt.ylabel('Count') plt.show()

#regionwise plot sns.countplot(x = 'region', data = df,hue=df.region) plt.title('Region Distribution') plt.xlabel('Region') plt.ylabel('Count') plt.show()

#count of smokers sns.countplot(x = 'smoker', data = df, hue=df.smoker) plt.title('Smoker Count') plt.xlabel('Smoker') plt.ylabel('Count') plt.show()

#charges distribution sns.histplot(df.charges,bins=20, kde=True,color='red') plt.title('Charges Distribution') plt.xlabel('Medical Expense') plt.ylabel('Count') plt.show()

#plotting the coorelation heatmap plt.figure(figsize=(10,10)) sns.heatmap(df.corr(),annot=True,cmap='coolwarm') plt.show()

sns.catplot(x="smoker", kind="count",hue = 'sex', data=df) plt.title('Smoker Count with gender') plt.xlabel('Smoker') plt.ylabel('Count') plt.show()

plt.figure(figsize=(12,5)) plt.title("Box plot for charges of women") sns.boxplot(y="smoker", x="charges", data = df[(df.sex == 0)] , orient="h")

plt.figure(figsize=(12,5)) plt.title("Box plot for charges of men") sns.boxplot(y="smoker", x="charges", data = df[(df.sex == 1)] , orient="h")

#non smokers charge distribution plt.figure(figsize=(7,5)) plt.title("scatterplot for charges of non smokers") sns.scatterplot(x="age", y="charges", data = df[(df.smoker == 0)]) plt.xlabel('Age') plt.ylabel('Medical Expense') plt.show()

#smokers charge distribution plt.figure(figsize=(7,5)) plt.title("scatterplot for charges of smokers") sns.scatterplot(x="age", y="charges", data = df[(df.smoker == 1)]) plt.xlabel('Age') plt.ylabel('Medical Expense') plt.show()

#age charges distribution sns.lmplot(x="age", y="charges", data = df, hue = 'smoker') plt.xlabel('Age') plt.ylabel('Medical Expense') plt.show()

#bmi charges distribution for obese people plt.figure(figsize=(7,5)) sns.displot(df[(df.bmi >= 30)]['charges']) plt.title('Charges Distribution for Obese People') plt.xlabel('Medical Expense') plt.show()

plt.figure(figsize=(7,5)) sns.displot(df[(df.bmi < 30)]['charges']) plt.title('Charges Distribution for Non Obese People') plt.xlabel('Medical Expense') plt.show()

x_train, x_test, y_train, y_test = train_test_split(df.drop('charges',axis=1), df['charges'], test_size=0.2, random_state=0)

print(x_train.shape) print(y_train.shape) print(x_test.shape) print(y_test.shape)

lr=LinearRegression() lr

#model training lr.fit(x_train,y_train) lr.score(x_train,y_train)

y_pred=lr.predict(x_test)

from sklearn.preprocessing import PolynomialFeatures poly_reg=PolynomialFeatures(degree=2) poly_reg

#transforming the features to higher degree x_train_poly = poly_reg.fit_transform(x_train) #splitting the data x_train, x_test, y_train, y_test = train_test_split(x_train_poly, y_train, test_size=0.2, random_state=0)

plr = LinearRegression() #model training plr.fit(x_train,y_train) #model accuracy plr.score(x_train,y_train)

#model prediction y_pred = plr.predict(x_test)

#decision tree regressor from sklearn.tree import DecisionTreeRegressor dtree = DecisionTreeRegressor() dtree

#model training dtree.fit(x_train,y_train) #model accuracy dtree.score(x_train,y_train)

#model prediction dtree_pred = dtree.predict(x_test)

from sklearn.metrics import mean_squared_error,mean_absolute_error,r2_score

#distribution of actual and predicted values plt.figure(figsize=(7,5)) ax1 = sns.distplot(y_test,hist=False,color='r',label='Actual Value') sns.distplot(y_pred,hist=False,color='b',label='Predicted Value',ax=ax1) plt.title('Actual vs Predicted Values for Linear Regression') plt.xlabel('Medical Expense') plt.show()

print('MAE:', mean_absolute_error(y_test, y_pred)) print('MSE:', mean_squared_error(y_test, y_pred)) print('RMSE:', np.sqrt(mean_squared_error(y_test, y_pred))) print('R2 Score:', r2_score(y_test, y_pred))

#acutal vs predicted values for polynomial regression plt.figure(figsize=(7,5)) ax1 = sns.distplot(y_test,hist=False,color='r',label='Actual Value') sns.distplot(y_pred,hist=False,color='b',label='Predicted Value',ax=ax1) plt.title('Actual vs Predicted Values for Polynomial Regression') plt.xlabel('Medical Expense') plt.show()

#distribution plot of actual and predicted values plt.figure(figsize=(7,5)) ax = sns.distplot(y_test, hist=False, color="r", label="Actual Value") sns.distplot(dtree_pred, hist=False, color="b", label="Fitted Values" , ax=ax) plt.title('Actual vs Fitted Values for Decision Tree Regression') plt.xlabel('Medical Expense') plt.ylabel('Distribution') plt.show()

print('MAE:', mean_absolute_error(y_test, dtree_pred)) print('MSE:', mean_squared_error(y_test, dtree_pred)) print('RMSE:', np.sqrt(mean_squared_error(y_test, dtree_pred))) print('Accuracy:', dtree.score(x_test,y_test))

df.head()