Quick neural network model for predicting Titanic survival

requirements.txt

keras
tensorflow
pandas
xlrd
h5py

test_titanic.py

import keras
import numpy

model = keras.models.load_model('titanic.hdf')

input_data = [1, 30, 1, 70, 1, 0, 0, 0, 1]
input_array = numpy.array(input_data).reshape(1, 9)
result = model.predict(input_array)

print(result[0,0])

train_titanic.py

import pandas
import numpy
import keras
from keras.layers import Dense, Dropout

data = pandas.read_excel('titanic3.xls')
data['Age'] = data['Age'].fillna(data['Age'].mean())
data['Fare'] = data['Fare'].fillna(data['Fare'].mean())
data['Family_size'] = data['SibSpo'] + data['Parents'] + 1
data['Fare_per_person'] = data['Fare'] / data['Family_size']

training_data = []
target_data = []

for row in data.itertuples():
	item = []
	item.append(row.Pclass)
	item.append(row.Age)
	item.append(row.Family_size)
	item.append(row.Fare_per_person)
	item.append(row.Embarked=='Q')
	item.append(row.Embarked=='S')
	item.append(row.Embarked=='C')
	item.append(row.Sex=='male')
	item.append(row.Sex=='female')
	training_data.append(item)
	target_data.append([row.Survived])

training_data_in = numpy.array(training_data)
target_data_in = numpy.array(target_data)

model = keras.Sequential()
model.add(Dense(15, activation='relu', input_dim=9))
model.add(Dropout(0.2))
model.add(Dense(10, activation='relu'))
model.add(Dropout(0.2))
model.add(Dense(10, activation='relu'))
model.add(Dense(1, activation='sigmoid'))
model.compile(optimizer="adam", loss="binary_crossentropy", metrics=['accuracy'])
model.summary()

model.fit(training_data_in, target_data_in, epochs=100, batch_size=200, validation_split=0.2)

model.save('titanic.hdf')