Reference no: EM132143654
Prepare a 3 - 10 pages long report
Prepare a presentation with 5 - 9 slides. The slides will include introduction (need and similar work), theoretical background (tested neural networks), Data, Results (Comparison of the performances of the tested neural networks, demonstrate the performance of your best neural network), Conclusion
Please give
1. a hard copy of the report
2. send the report and the PowerPoint presentation to
Select a modeling and prediction problem. Work with actual data. Separate part of the data for training and others for testing. Train neural networks with the training data to represent it. (Data collection and reading it with Matlab is 30 points).
Teach it to
- 3 different neural network training algorithms - Select the best one
- try at least 3 different number of nodes at the hidden layer - Select the best case
- Evaluate the performance of the network on the test data the best trained NN has never seen before.
Outline the system you select and performance of the network in the report. You will present the result in the class with a Power Point presentation (report and presentation - 30 points)
THEORETICAL BACKGROUND
- Daily Sunrise times have been "trained" in Los Angeles for the next 7 weeks based on the 7 weekdays
- Prior to the "Trained" days, the Previous 7 weeks have been "tested'
- The Neural Network will be able to Predict the upcoming times of sunrises
- The algorithms used in neural networks can rapidly calculates many different possibilities