adrigrillo/dl

Deep Learning

dl's Issues

Assignment 2

Using the data contained in the file Xtrain.mat the objective is to train a model to predict the following step, that will be given the test day.

Characteristics

• The test length is of 200 time steps, therefore the model has to predict 200 steps.
• The points predicted has to be added to the considered windows, to predict the following.

Assigment 1

Assignment 1:

Download the Iris data and explore it.

• Analyse and visualize the dataset. How many classes, instances, features, etc...

Select feature number 0 and 2 and the first samples and do the following analysis:

• Implement the logistic regression model that discussed in the class (with and
  without regularization). Use Gradient descent algorithm for updating the parameters.
  Plot the results.
• Explore the influence of learning rate on the convergence of the model.
  Tune the learning rate.
• Repeat steps 1-4 for different randomly selected features (e.g. 1 and 3 or 2
  and 3. and compare the results.
• Give the option to the users to select whatever combination of features they
  want and your code will do the rest.
• Try your model on Monk2 dataset and report the results, (test accuracy,
  training accuracy, optimal learning rate, loss value... Use the last 20% of the
  data as test.