I designed a neural network for binary classification in MATLAB R2015a. Now I want these results:
* What are differences between two classes? How system detects a sample is from class 1 or 2? For example some if,then functions or some ranges. Something which can help user to have clearer view of the results. For example neural network found that a sample is from class 1. I should say in report that what are differences between in this sample and other samples (class 2) for this results (based on neural network inputs). I know this is a black box but I need more results. I'm checking the trained system using out-of-sample data.
* I want know the effects of inputs on output. Which input (feature) is more important (has higher weight) on output of trained neural network.
Thanks.
For both these questions you will have to examine the inner workings of your model.i.e. the weights connecting layers and the activation energies that are output from the neurons.
1. Possibly clamp one (or multiple) input vectors from each class and examine the different energies in calculated in the hidden layers and see what different neurons are firing, determining which features are most relevant to each class.
2. Visualise or examine the weights coming from the input vector, would probably give you a good idea of this.
Thank you Jim. Can you describe more about 1? Suppose that i have 5 inputs (features) that i checked we have higher accuracies with these input's combination. So now i want find class 1 and class 2 characteristics. My neural network has two hidden layers. First one has 5 neurons and second one has 3 neurons. My hidden layers transfer function is 'tansig' and output transfer function is 'softmax' for reporting probabilities for outputs. Now what should i do ?
Dear Erm Erm,
Look the link, may be useful.
chrome-extension://oemmndcbldboiebfnladdacbdfmadadm/http://papers.nips.cc/paper/853-classification-of-multi-spectral-pixels-by-the-binary-diamond-neural-network.pdf
http://www.webpages.ttu.edu/dleverin/neural_network/neural_networks.html
Regards, Shafagat
Ern,
Neural networks have all good qualities of classifier expect one i.e. reasoning. The classifier structure is non-trivial for human understanding. Try decision tree for understanding. You may use NN for deployment, but DT for gaining insights into the decision making process.
Best wishes
You can try to analyze the classification output when varying the inputs. This method is called "sensitivity analysis". When, given a set of values (exploration point) in all input variables except in one, you can vary this excluded variable, and analyze the classification output, so you should be able to see when output changes for certain input values.
- Badges
- Science topic