how to explain this line? Ch4,04_svm_kernels.py,line71:prediction = tf.sign(prediction_output - tf.reduce_mean(prediction_output))

[qwer10]

Ch4,04_svm_kernels.py,line71:
prediction = tf.sign(prediction_output - tf.reduce_mean(prediction_output))
I can't find that line in any book/posts.
Instead, my book said when you predict, you should calculate the bias first.
Who can explain that line? Thank you very much!