- Can use single neuron to do logistic regression (use logit as a smooth function)
- Use python’s Kera package (keras.io)
- Challenge model with a small neural network of two connected hidden layers with each 10 neurons
- Identify possible non linearities by looking at default rate occurrence per driver
- Examine gini as a function of driver threshold
- For bigger datasets use Python Theano or Tensorflow
Neural network as challenger in classification tasks
Original by DataQuest, 2017, 6 pages
This summary note was posted on 21 July 2018, by Reinie
in Credit risk
Latest Hamster Notes
- Measure what Matters posted in Management
- PSPO I posted in Agile
- Stuff on Scrum posted in Agile
- 3 tips to create a courageous space posted in Management
- The Lean Strategy posted in Management Personal Development
- 6 traits of an inclusive leader posted in Management
- Myers Briggs Type Indicator posted in Personal Development
- Positive Influence posted in Management
- Start with Why posted in Management
- 4 steps to optimise product value posted in Agile Management