Data Analytics Foundations

In this assignment you will build and tune a deep neural network.
1. Select a problem for which there is an open source dataset or for which you have data from other sources.
There must be several independent variables. If you have been working with time series data, you need to engineer features from dates and times.
2. Select deep learning architecture suitable for your problem. Consider different network architectures for your problem.
3. Tune the selected network(s). At a minimum, tune the number of layers, number of neurons, and two other parameters. Depending on your problem, you should consider tuning additional parameters. Some of the networks you consider should have at least five hidden layers with exception of RNN for which three hidden layers are sufficient.
4. Assess the accuracy of the tuned deep learning model(s) and compare its accuracy with accuracy .. If you are using a different data set, compare the tuned neural network with another model (does not have to be tuned) of your choice.
————————————————————————————————————————————
Descriiption of the forecasting problem and data.
• Overview of the network architecture(s) you have used including important parameters that affect the network operation. Discuss the reasons for selecting a specific architecture. Make sure you include parameters you will tune.
• Descriiption of the process you have used including data pre-processing, feature generation, model training/testing, and evaluation.
• Results. This should include accuracy measures (more than one) achieved through the training process, graphs demonstrating final accuracy as well as the accuracy through the tuning process.
• Code. Although there are no marks for the code itself, marks will be deducted if the code does not match the rest of the report.

IMPORTANT: The report itself should not contain any code. The code is submitted as an appendix or a separate file. The report should describe what you have done in general terms – the same process and algorithms could be implemented in different languages.