Reinforcement Learning with TensorFlow
上QQ阅读APP看书,第一时间看更新

Notation

Let the data be of the form , where:

  • , (number of classes = 2 because it's a binary classification)

  • is 'n' dimensional, that is, (refers to the preceding diagram)

  • The number of training examples is m. Thus the training set looks as follows:

    • .

    • m = size of training dataset.

    • And, since , where, each .

    • Therefore, is a matrix of size n * m, that is, number of features * number of training examples.

    • , a vector of m outputs, where, each .

    • Parameters : Weights , and bias , where and is a scalar value.