We build a neural network to predict the value of a financial indicator, when running the neural network on the same inputs multiple times, each time we receive different network with different result. We seek a method/algorithm to choose the “best networks” that will minimize the prediction error. Full description of the problem in the attached PDF file We also attached some data file in XLS format.
5 freelance font une offre moyenne de $472 pour ce travail
Hi I'm an electrical engineer with excellent experience in Neural Network (NN) technique I've gone through the attached files and feel confident to help you in your project to find and build an appropriate algorithm Plus
hi! neural networks can be tricky. the reason that you get a "different" network every time is that the backpropagation algorithm that optimises the network starts with random weights across the network. so every time Plus
Hi, I have hands-on experience with machine learning techniques. I have been in the area for more than 3 years. I can analyze the data for you and help you with practical advices on applying machine learning algorithm Plus