Multi-class classification

Introduction to Deep Learning with Keras

Miguel Esteban

Data Scientist & Founder

Throwing darts

The dataset

The architecture

The output layer

Multi-class model

# Instantiate a sequential model
# ...

# Add an input and hidden layer
# ...

# Add more hidden layers
# ...

# Add your output layer
model.add(Dense(4, activation='softmax')

Categorical cross-entropy

model.compile(optimizer='adam', loss='categorical_crossentropy')

Preparing a dataset

import pandas as pd
from tensorflow.keras.utils import to_categorical

# Load dataset
df = pd.read_csv('data.csv')

# Turn response variable into labeled codes
df.response = pd.Categorical(df.response)
df.response = df.response.cat.codes

# Turn response variable into one-hot response vector
y = to_categorical(df.response)

One-hot encoding

Let's practice!

Introduction to Deep Learning with Keras