Please I need some helpe to implement this codes:

# Implementation of the perceptron learning algorithm. Support the pocket version for linearly unseparatable data.

#Important observation:

# - The PLA can increase or decrease $w[0]$ by 1 per update, so if there is a big difference between $w^*[0]$ and the #initial value of $w[0]$, the PLA is likely to take a long time before it halts. However, the theoretical bound $O((L/d)^2)$ #step of course still holds, where $L = \max\{\lVert x Vert\}$ and $d$ is the margine size.

# - This can solved by always have feature values within [0,1], because by doing so, the $x_0=1$ becomes relatively larger (or one can also say $x_0$ becomes fairly as important as other feathers), which makes the changes to $w[0]$ much faster. This is partially why nueral network requires all feature value to be [0,1] --- the so-called data normalization process!!!

# Another reason for normalizing the feature into [0,1] is: no matter which Z space the samples are tranformed to, the Z-space sample features will still be in the [0,1] range.

import numpy as np

#import sys

#sys.path.append("..")

from utils import MyUtils

class PLA:

def __init__(self, degree=1):

self.w = None

self.degree = degree

def fit(self, X, y, pocket = True, epochs = 100):

''' find the classifer weight vector and save it in self.w

X: n x d matrix, i.e., the bias feature is not included.

It is assumed that X is already normalized be data preprocessing.

y: n x 1 vector of {+1, -1}

degree: the degree of the Z space

return self.w

'''

if(self.degree > 1):

X = MyUtils.z_transform(X, degree=self.degree)

### BEGIN YOUR SOLUTION

raise NotImplementedError()

### END YOUR SOLUTION

return self.w

def predict(self, X):

''' x: n x d matrix, i.e., the bias feature is not included.

return: n x 1 vector, the labels of samples in X

'''

if(self.degree > 1):

X = MyUtils.z_transform(X, degree = self.degree)

### BEGIN YOUR SOLUTION

raise NotImplementedError()

### END YOUR SOLUTION

def error(self, X, y):

''' X: n x d matrix, i.e., the bias feature is not included.

y: n x 1 vector

return the number of misclassifed elements in X using self.w

'''

if(self.degree > 1):

X = MyUtils.z_transform(X, degree = self.degree)

### BEGIN YOUR SOLUTION

raise NotImplementedError()

### END YOUR SOLUTION