Problem $2$:

Load the data set Q consisting of a single explanatory variable $\backslash (\backslash$boldsymbol$\{$x$\}\backslash )$ and dependent variable $\backslash (\backslash$boldsymbol$\{$y$\}\backslash ).$

$1.$ Shuffle the rows of $\backslash ($ Q $\backslash )$ and separate the data into a test set and training set. Use $\backslash (30\backslash \%\backslash )$ of the data for the testing set and the remaining $70\backslash \%$ for training set.

$[42]$: Q $=$ np$.$genfromtxt$(\text{'}$dataLS$.$txt$\text{'},$delimiter$=\text{'},\text{'})$

np$.$random.shuffle$\backslash$#$\backslash$#$\backslash$#$\backslash$# use polyfit

$2.$ Build $4$ different models on the training set, each of different degree polynomials $\backslash (\backslash$boldsymbol$\{$p$\}=1,2,3,4\backslash ).$ With each model:

a$.$ Print a graph of the data with a graph of the model. Use blue for training data, red for the testing data and use black for the model.

b$.$ Below the graph, print the model parameters and the mean of the squared error on the testing set.