Naive Bayes model for weight and gender data

nb.py

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline
import seaborn as sns


actid = pd.read_csv("https://raw.githubusercontent.com/Thinkful-Ed/curric-data-001-data-sets/master/ideal-weight/ideal_weight.csv")

actid.head()

# Remove '' from headers
actid.columns = [name.strip("''") for name in actid.columns.values.tolist()]

# Remove '' from sex column
actid['sex'] = actid['sex'].map(lambda x: x.strip("''"))

# Plot ideal and actual weights
plt.hist(actid['ideal'], bins=28, alpha=0.5, label='Ideal')
plt.hist(actid['actual'], bins=28, alpha=0.5, label='Actual')
plt.legend(loc='upper right')
plt.show()

difference = actid['actual'] - actid['ideal']
plt.hist(difference, bins=28)
plt.show()

# Convert 'sex' column to categorical
actid['sex'] = pd.Categorical(actid['sex']).codes

# Male is now 1 and Female 0. (More females than males in the dataset)
print float(sum(actid['sex'])) / float(len(actid['sex']))


# Train test split
from sklearn.cross_validation import train_test_split

y = actid.sex
y = np.array(y)

X = actid[['actual', 'ideal', 'diff']]
X = np.array(X)

X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size=.30, random_state=42)

# Build NaiveBayes Model
from sklearn.naive_bayes import GaussianNB

clf = GaussianNB()

clf = GaussianNB()
clf.fit(X_train,y_train)

y_pred = clf.predict(X_test)

print sum(y_test != y_pred)

# Predict new round of weights. (Prediction is Male)
first_trial = clf.predict([[145, 160, -15]])

# Predict second round of weights (Prediction is Female)
second_trial = clf.predict([[160, 145, 15]])