कस्टम loss functions

Python में Finance के लिए Machine Learning

Nathan George

Data Science Professor

दिशा का मेल नहीं

दिशात्मक दंड के साथ MSE

यदि prediction और target की दिशा मिलती है:

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

यदि नहीं मिलती:

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

कस्टम loss functions लागू करना

import tensorflow as tf

फंक्शन बनाना

import tensorflow as tf

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

Mean squared error loss

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)

Keras में कस्टम loss जोड़ें

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)

सही दिशा की जाँच

tf.less(y_true * y_pred, 0)

सही दिशा:

neg * neg = pos
pos * pos = pos

गलत दिशा:

neg * pos = neg
pos * neg = neg

tf.where() का उपयोग

# 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))

सबको जोड़ना

# 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

कस्टम loss का उपयोग

# 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)

बो-टाई आकार

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()

बो-टाई प्लॉट

अपना loss function बनाएँ!

Python में Finance के लिए Machine Learning