packtpublishing/machine-learning-for-finance

Machine Learning for Finance, published by Packt

machine-learning-for-finance's Issues

TypeError: only integer scalar arrays can be converted to a scalar index

I keep getting this error, trying to solve it but havent be able to, I am using python 3.0 from mini-conda, thanks.

epoch = 5000 # Number of games played in training, I found the model needs about 4,000 games till it plays well

# Train the model
# For simplicity of the noteb
hist = train(model, epoch, verbose=1)

print("Training done")

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-11-1e110d42c25b> in <module>
      3 # Train the model
      4 # For simplicity of the noteb
----> 5 hist = train(model, epoch, verbose=1)
      6 
      7 print("Training done")

<ipython-input-10-b5900a520e40> in train(model, epochs, verbose)
     12         game_over = False
     13         # get initial input
---> 14         input_t = env.observe()
     15 
     16         while not game_over:

<ipython-input-2-a9f37a9957fc> in observe(self)
     57 
     58     def observe(self):
---> 59         canvas = self._draw_state()
     60         return canvas.reshape((1, -1))
     61 

<ipython-input-2-a9f37a9957fc> in _draw_state(self)
     36         canvas[state[0], state[1]] = 1  # draw fruit
     37 
---> 38         canvas[-1, state[2] -1:state[2] + 2] = 1  # draw basket
     39 
     40         return canvas

TypeError: only integer scalar arrays can be converted to a scalar index

NameError: name 'pred_func' is not defined

@JannesKlaas
Hi Jannes,
In chapater 1,
...
A deeper network
...

Helper function to plot a decision boundary.

If you don't fully understand this function don't worry, it just generates the contour plot below.

def plot_decision_boundary(pred_func):
# Set min and max values and give it some padding
x_min, x_max = X[:, 0].min() - .5, X[:, 0].max() + .5
y_min, y_max = X[:, 1].min() - .5, X[:, 1].max() + .5
h = 0.01
# Generate a grid of points with distance h between them
xx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))
# Predict the function value for the whole gid
Z = pred_func(np.c_[xx.ravel(), yy.ravel()])
Z = Z.reshape(xx.shape)
# Plot the contour and training examples
plt.contourf(xx, yy, Z, cmap=plt.cm.Spectral)
plt.scatter(X[:, 0], X[:, 1], c=y.flatten(), cmap=plt.cm.Spectral)

---

NameError Traceback (most recent call last)
/tmp/ipykernel_27/694468409.py in
15
16 #Plot the decision boundary
---> 17 plot_decision_boundary(lambda x: predict(model, x))
18 plt.title("Decision Boundary for hidden layer size 3")
19

/tmp/ipykernel_27/2764471563.py in plot_decision_boundary(pre_func)
114 xx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))
115 #Predict the function value for the whole gid
--> 116 Z = pred_func(np.c_[xx_ravel(), yy_ravel()])
117 Z = Z.reshape(xx.shape)
118 #Plot the contour and training examples

NameError: name 'pred_func' is not defined

ImportError: cannot import name 'SGD' from 'keras.optimizers'

Hi,

And even with the help of SOF I still cannot reproduce the numbers you've got in your git. I'm getting the following:

Epoch 1/5
13013/13013 [==============================] - 17s 1ms/step - loss: 825128.0000 - acc: 0.8998 - val_loss: 211068.5156 - val_acc: 0.9019
Epoch 2/5
13013/13013 [==============================] - 16s 1ms/step - loss: 766153.7500 - acc: 0.9068 - val_loss: 582742.2500 - val_acc: 0.8027

---

ImportError Traceback (most recent call last)
in ()
1 from keras.models import Sequential
2 from keras.layers import Dense, Activation
----> 3 from keras.optimizers import SGD

ImportError: cannot import name 'SGD' from 'keras.optimizers' (/usr/local/lib/python3.7/dist-packages/keras/optimizers.py)

---

NOTE: If your import is failing due to a missing package, you can
manually install dependencies using either !pip or !apt.

To view examples of installing some common dependencies, click the
"Open Examples" button below.

Answers for the exercises

@JannesKlaas
Hi Jannes,

Thank you for the amazing book and the tutorials, I really appreciate the time and effort you have put into the book.

I am wondering if there is somewhere that you have posted the answers of the exercieses for us to double check?
Since this is completely self-study, having answers to look up would be helpful to let me know that I am at least in the right direction.

Thanks.

At page 54 , investigate zero balances as a feature of fraud detection

Hi,

At page 54 of this book, there is a formula for counting origin accounts have insufficient funds. As the code below:

     len(dfOdd[(dfOdd.oldBalanceOrig <= dfOdd.amount)]) / len(dfOdd)

This formula get 0.89664... , close to 90% of the odd transactions have insufficient funds.
Since we want accounts of insufficient funds before transaction. why use " <= " not. " < " ?
( The latter one get 0.21762... )

Regards,
Steven