Scribd
Upload
12 views34 pages

MTH408 Machine - Learning - ANN

The document discusses neural networks and their applications in finance. It covers topics like logistic units, multiple output units, neural network classification, representing Boolean functions, the perceptron learning rule, learning in neural networks, forward and backpropagation, training neural networks via gradient descent with backpropagation, random initialization of weights, putting all together to train a neural network, and popular neural network types.

Uploaded by

junzi2000
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
12 views34 pages

MTH408 Machine - Learning - ANN

The document discusses neural networks and their applications in finance. It covers topics like logistic units, multiple output units, neural network classification, representing Boolean functions, the perceptron learning rule, learning in neural networks, forward and backpropagation, training neural networks via gradient descent with backpropagation, random initialization of weights, putting all together to train a neural network, and popular neural network types.

Uploaded by

junzi2000
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 34

Dr.

David Liu
Department of Financial and Actuarial Mathematics
School of Science, Xi’an Jiaotong-Liverpool University

1
2
3
4
Neuron model - Logistic unit

5
6
+
+

7
8
s = [s0, s1, s2, s3]

9
Multiple output units (multiple classes)

10
Neural network classification

11
12
Representing Boolean functions

13
Perceptron learning rule

14
Learning in Neural Network

15
16
17
m

2m

18
Forward propagation

19
Backpropagation intuition

20
Backpropagation intuition

21
Backpropagation intuition

22
Backpropagation intuition

23
Backpropagation intuition

24
Backpropagation intuition

25
Training neural network via gradient Descent with backpropagation

m m

26
27
28
29
Random initialization of weights

30
Random initialization of weights

31
Putting all together – Training a neural network

32
Putting all together – Training a neural network

Step 5 is not really a


must. Theoretical part is
not provided here.
So our program doesn’t
implement this step.

33
Popular neural networks

BP NN is the most popular and fundamental network used for financial problems.

1. Hybrid Neural networks for financial options (an article for your reading)
2. GRNN
3. LSTM
4. Convolutional NN
5. …..

34

You might also like