Algorithm Discussion

$$ Esci mark this page

Project due Jun $26,202417$:$29$ IST

cemp$-$iwe $,-$

Once you have completed the implementation of the $3$ learning algorithms, you should qualitatively verify your implementations. In

main.py we have included a block of code that you should uncomment. This code loads a $20$ dataset from toy$\_$data.txt$,$ and trains your models using $T=10,=0\mathrm{.}2$

main.py will compute $$ and ${}_0$ for each of the learning algorithms that you have written. Then, it will call plot$\_$toy$\_$data to plot the resulting model and boundary.

Plots

$\frac{0\mathrm{.}5}{6}$ points $($graded$)$

In order to verify your plots, please enter the values of $$ and ${}_0$ for all three algorithms.

$($For example, if $=(1,0\mathrm{.}5),$ then type $1,0\mathrm{.}5$ without the brackets. Make sure your answers are correct up to $4$ decimal places.$)$

For the perceptron algorithm:

$=$

$1\mathrm{.}8809$

${}_0=$

For the average perceptron algorithm:

$=$

$3\mathrm{.}5092,0\mathrm{.}9556$

${}_0=x^{**}$

For the Pegasos algorithm:

$=$

$0\mathrm{.}1,0\mathrm{.}1$

${}_0=$

$$