Aangepaste verliesfuncties

Machine Learning voor finance in Python

Nathan George

Data Science Professor

richting komt niet overeen

MSE met richtingsstraf

Als de richting van voorspelling en target overeenkomt:

$ \sum(y - \hat{y})^2$

Zo niet:

$\sum (y - (\hat{y}))^2 * \text{penalty} $

Aangepaste verliesfuncties implementeren

import tensorflow as tf

Een functie maken

import tensorflow as tf

# create loss function
def mean_squared_error(y_true, y_pred):

Mean squared error-verlies

import tensorflow as tf

# create loss function
def mean_squared_error(y_true, y_pred):

    loss = tf.square(y_true - y_pred)
    return tf.reduce_mean(loss, axis=-1)

Aangepaste verliesfunctie aan keras toevoegen

import tensorflow as tf

# create loss function
def mean_squared_error(y_true, y_pred):
    loss = tf.square(y_true - y_pred)
    return tf.reduce_mean(loss, axis=-1)

# enable use of loss with keras
import keras.losses
keras.losses.mean_squared_error = mean_squared_error

# fit the model with our mse loss function
model.compile(optimizer='adam', loss=mean_squared_error)
history = model.fit(scaled_train_features, train_targets, epochs=50)

Controleren op juiste richting

tf.less(y_true * y_pred, 0)

Juiste richting:

neg * neg = pos
pos * pos = pos

Verkeerde richting:

neg * pos = neg
pos * neg = neg

tf.where() gebruiken

# create loss function
def sign_penalty(y_true, y_pred):
    penalty = 10.
    loss = tf.where(tf.less(y_true * y_pred, 0), 
                    penalty * tf.square(y_true - y_pred), 
                    tf.square(y_true - y_pred))

Alles samenbrengen

# create loss function
def sign_penalty(y_true, y_pred):
    penalty = 100.
    loss = tf.where(tf.less(y_true * y_pred, 0),
                    penalty * tf.square(y_true - y_pred), 
                    tf.square(y_true - y_pred))

    return tf.reduce_mean(loss, axis=-1)

# enable use of loss with keras
keras.losses.sign_penalty = sign_penalty

De aangepaste verliesfunctie gebruiken

# create the model
model = Sequential()
model.add(Dense(50,
                input_dim=scaled_train_features.shape[1],
                activation='relu'))
model.add(Dense(10, activation='relu'))
model.add(Dense(1, activation='linear'))

# fit the model with our custom 'sign_penalty' loss function
model.compile(optimizer='adam', loss=sign_penalty)
history = model.fit(scaled_train_features, train_targets, epochs=50)

De vlinderdas-vorm

train_preds = model.predict(scaled_train_features)
# scatter the predictions vs actual
plt.scatter(train_preds, train_targets)
plt.xlabel('predictions')
plt.ylabel('actual')
plt.show()

vlinderdas-plot

Maak je eigen verliesfunctie!

Machine Learning voor finance in Python