How do I improve my Python code and is it correct? If not, can you help me fix it$?$

Is this the best outcome? I got an $86\mathrm{.}67$ which I assume is $86\%$ accuracy. Please only edit steps $2,3,5,6,7,8,9$ with the ### YOUR CODE HERE label. Thank you.

Step $1$

Import the libraries.

import numpy as np

from csv import reader

from random import seed

from random import randrange

Step $2$

Load the csv file.

def load$\_$csv$($filename$,$ skip $=$ False$)$:

dataset $=$ list$()$

### YOUR CODE HERE

# Opens the file in read$-$only mode

with open$($filename$,\text{'}$r$\text{'})$ as file: # Open the file

csv$\_$reader $=$ reader$($file$)$

if skip:

next$($csv$\_$reader$)$ # Skip header

for row in csv$\_$reader:

if not row:

continue

dataset.append$([$float$($value$)$ for value in row$])$ # Convert values to float

$$

return dataset

Step $3$

Split the dataset into X$\_$train, Y$\_$train, X$\_$test, Y$\_$test sets.

def train$\_$test$\_$split$($dataset$,$ split$)$:

### YOUR CODE HERE

# Shuffle and split the dataset

np$.$random.shuffle$($dataset$)$

train$\_$size $=$ int$($split $*$ len$($dataset$))$

train$\_$data $=$ dataset$[$:train$\_$size$]$

test$\_$data $=$ dataset$[$train$\_$size:$]$

$$

# Split features $($X$)$ and labels $($y$)$

X$\_$train $=[$row$[$:$-1]$ for row in train$\_$data$]$

y$\_$train $=[$row$[-1]$ for row in train$\_$data$]$

X$\_$test $=[$row$[$:$-1]$ for row in test$\_$data$]$

y$\_$test $=[$row$[-1]$ for row in test$\_$data$]$

$$

return X$\_$train, y$\_$train, X$\_$test, y$\_$test

Step $4$

Defining the Perceptron class that contains the weights, bias, learning rate and epochs.

class Perceptron:

def $\_\_$init$\_\_($self$,$ input$\_$size, bias, learning$\_$rate, epochs$)$:

self.weights $=$ np$.$zeros$($input$\_$size$)$

self.bias $=$ bias

self.learning$\_$rate $=$ learning$\_$rate

self.epochs $=$ epochs

Step $5$

Define the activation function.

def activation$\_$function$($x$)$:

# Step function for perceptron

### YOUR CODE HERE

result $=1$ if x $>=0$ else $0$ # Use a step function

return result

Step $6$

Defining the predict function with the inputs, weights and bias values.

def predict$($inputs$,$ weights, bias$)$:

### YOUR CODE HERE

weighted$\_$sum $=$ np$.$dot$($inputs$,$ weights$)+$ bias

return activation$\_$function$($weighted$\_$sum$)$

Step $7$

Define the train function.

def train$($X$\_$train, y$\_$train, learning$\_$rate, epochs, weights, bias$)$:

prediction $=$ None

error $=$ None

for $\_$ in range$($epochs$)$:

### YOUR CODE HERE

for i in range$($len$($X$\_$train$))$:

prediction $=$ predict$($X$\_$train$[$i$],$ weights, bias$)$ # Calculate prediction

error $=$ y$\_$train$[$i$]-$ prediction # Calculate the error

weights $+=$ learning$\_$rate $*$ error $*$ np$.$array$($X$\_$train$[$i$])$ # Update weights

bias $+=$ learning$\_$rate $*$ error # Update bias

$$

return weights, bias

Step $8$

Define the accuracy for the perceptron.

def perceptron$\_$accuracy$($y$,$ y$\_$hat$)$:

# overwrite the accuracy value with your own code

accuracy $=0$

### YOUR CODE HERE

correct $=$ sum$(1$ for actual, predicted in zip$($y$,$ y$\_$hat$)$ if actual $==$ predicted$)$

accuracy $=$ correct $/$ len$($y$)*100$

return accuracy

Step $9$

Implemented the Perceptron Nerual Network.

# Set the seed

seed$(1)$

$$

# Load the csv file

$$

filename $=$ 'moons.csv$\text{'}$

dataset $=$ load$\_$csv$($filename$,$ skip$=$True$)$

$$

# Configure the perception with the bias, learning rate and epochs

$$

# Note the initial values are dummy and must changed for an accurate network

$$

# The split value for the training and test sets

custom$\_$split $=0$

$$

# The bias term is a constant value added to the weighted sum of inputs

custom$\_$bias $=-1$

$$

# The learning rate controls how much the weights are adjusted during training

custom$\_$learning$\_$rate $=-1$

$$

# The number of epochs defines how many times the perceptron will iterate over the training data

custom$\_$epochs $=-1$

$$

# Set your values here

$$

###

### YOUR CODE HERE

custom$\_$split $=0\mathrm{.}8$ # Use $80\%$ of the data for training and $20\%$ for testing

custom$\_$bias $=0$ # Initial bias value

custom$\_$learning$\_$rate $=0\mathrm{.}01$ # Learning rate for adjusting weights

custom$\_$epochs $=100$ # Number of epochs to train the model

###

$$

# Split the dataset for both training and testing

$$

X$\_$train, y$\_$train, X$\_$test, y$\_$test $=$ train$\_$test$\_$split$($dataset$,$ split$=$custom$\_$split$)$

$$

perceptron $=$ Perceptron$($input$\_$size$=2,$ bias$=$custom$\_$bias, learning$\_$rate$=$custom$\_$learning$\_$rate, epochs$=$custom$\_$epochs$)$

$$

# Training

weights, bias $=$ train$($X$\_$train, y$\_$train, perceptron.learning$\_$rate, perceptron.epochs, perceptron.weights, perceptron.bias$)$

$$

# Predictions

y$\_$hat $=[]$

$$

# Testing

for i in range$($len$($X$\_$test$))$:

prediction $=$ predict$($X$\_$test$[$i$],$ weights, bias$)$

y$\_$hat.append$($prediction$)$

print$($f$"$Input: $\{$X$\_$test$[$i$]\},$ Predicted: $\{$prediction$\},$ Actual: $\{$y$\_$test$[$i$]\}")$

$$

# Test for Accuracy

perceptron$\_$accuracy$($y$\_$test, y$\_$hat$)$