Hello, I have the Codes on Jupyter Notebook with descriptions attached below I just need the data to be analyzed and following questions be answered: (Boston Housing)
Upload 1-predicting-house-prices.ipynb file to your Google Colab account and do the following:
1) In addition to recording the MAE metric during training, make a plot that shows the training error as well (put both curves in the same plot). Explain the behavior you see in the image.
2) Increase the size of the NN model (i.e., number of layers and/or number of neurons per layer) and study its effect on the training-testing loss figure.
3) Redo (1) but add L1 or L2 regularization on the fitted model. Change the regularization weights as well.
4) Study the effect of “number of folds” in your evaluations. What are the benefits/disadvantageous of using too many or too few folds?
5) Plot the histogram of the responses in the dataset. Would our NN modeling benefit from linear or nonlinear transformation of the output?
6) Fit the best NN model you can to the training data. Then, use it to predict the testing data (this testing should only be done once). Plot a scatter plot that shows your predictions vs. the target in the test data.
0 comments