Contact Us!
CoCalc Logo Icon
StoreFeaturesDocsShareSupportNewsAboutSign UpSign In
Avatar for stephanie's main branch.

Real-time collaboration for Jupyter Notebooks, Linux Terminals, LaTeX, VS Code, R IDE, and more,
all in one place. Commercial Alternative to JupyterHub.

| Download

"Guiding Future STEM Leaders through Innovative Research Training" ~ thinkingbeyond.education

Views: 1075
Image: ubuntu2204
Tweet
Share
Share
Kernel: Python 3
"""
First Code - No C Parameter, Only Gamma Parameter - Not Used for Poster
Using Digits Dataset of Scikit, Not MNIST Dataset
"""

import matplotlib.pyplot as plt
from sklearn import datasets, metrics, svm
from sklearn.model_selection import train_test_split

digits = datasets.load_digits()

"""
_, axes = plt.subplots(nrows = 1, ncols = 4, figsize = (10, 3))
for ax, image, label in zip(axes, digits.images, digits.target):
    ax.set_axis_off()
    ax.imshow(image, cmap = plt.cm.gray_r, interpolation = "nearest")
    ax.set_title("Training: %i" % label)
"""

n_samples = len(digits.images)
data = digits.images.reshape((n_samples, -1))
clf = svm.SVC(gamma = 0.001)

X_train, X_test, y_train, y_test = train_test_split(
    data, digits.target, test_size = 0.7, shuffle = False
)

clf.fit(X_train, y_train)
predicted = clf.predict(X_test)

"""
_, axes = plt.subplots(nrows=1, ncols=4, figsize=(10, 3))
for ax, image, prediction in zip(axes, X_test, predicted):
    ax.set_axis_off()
    image = image.reshape(8, 8)
    ax.imshow(image, cmap=plt.cm.gray_r, interpolation="nearest")
    ax.set_title(f"Prediction: {prediction}")
"""

print(
    f"Classification Report for Classifier {clf}:\n"
    f"{metrics.classification_report(y_test, predicted)}\n"
)
Classification Report for Classifier SVC(gamma=0.001): precision recall f1-score support 0 0.99 0.99 0.99 123 1 0.95 0.97 0.96 127 2 0.98 0.97 0.98 122 3 0.98 0.91 0.94 128 4 0.98 0.96 0.97 128 5 0.94 0.96 0.95 128 6 0.99 0.98 0.98 128 7 0.96 1.00 0.98 126 8 0.93 0.93 0.93 121 9 0.90 0.94 0.92 127 accuracy 0.96 1258 macro avg 0.96 0.96 0.96 1258 weighted avg 0.96 0.96 0.96 1258 
"""
Code using MNIST Dataset - 10,000 Images
Used Fixed Hyperparameters:
Gamma = 0.05
C = 1
Generated Classification Report and Confusion Matrix
"""

import numpy as np
import matplotlib.pyplot as plt
from sklearn.datasets import fetch_openml
from sklearn.model_selection import train_test_split
from sklearn import svm, metrics
from matplotlib.colors import Normalize

mnist = fetch_openml('mnist_784', version = 1, as_frame = False)

X = mnist.data
y = mnist.target.astype(int)

random_indices = np.random.choice(len(X), 10000, replace = False)
X_subset = X[random_indices]
y_subset = y[random_indices]

X_subset = X_subset / 255.0

X_train, X_test, y_train, y_test = train_test_split(X_subset, y_subset, test_size = 0.2, random_state = 42)

param_C = 1.0
param_gamma = 0.05
classifier = svm.SVC(C = param_C, gamma = param_gamma)

import datetime as dt
start_time = dt.datetime.now()
print(f"Start learning at {start_time}")

classifier.fit(X_train, y_train)

end_time = dt.datetime.now()
print(f"Stop learning at {end_time}")
elapsed_time = end_time - start_time
print(f"Elapsed learning time: {elapsed_time}")

predicted = classifier.predict(X_test)

print("Classification Report:\n", metrics.classification_report(y_test, predicted))
print("Accuracy:", metrics.accuracy_score(y_test, predicted))

cm = metrics.confusion_matrix(y_test, predicted)

class HighContrastNormalize(Normalize):
    def __init__(self, vmin = None, vmax = None, midpoint = None, clip = False):
        super().__init__(vmin, vmax, clip)
        self.midpoint = midpoint

    def __call__(self, value, clip = None):
        x, y = [self.vmin, self.midpoint, self.vmax], [0, 0.5, 1]
        return np.ma.masked_array(np.interp(value, x, y))

plt.figure(figsize = (10, 8))
plt.imshow(
    cm,
    interpolation = 'nearest',
    cmap = plt.cm.coolwarm,
    norm = HighContrastNormalize(vmin = 0, vmax = cm.max(), midpoint = cm.max() / 2)
)
plt.title("High-Contrast Confusion Matrix", fontsize = 16)
plt.colorbar()
classes = np.arange(10)
plt.xticks(classes, classes, fontsize = 12)
plt.yticks(classes, classes, fontsize = 12)
plt.ylabel("True Label", fontsize = 14)
plt.xlabel("Predicted Label", fontsize = 14)

thresh = cm.max() / 2
for i in range(cm.shape[0]):
    for j in range(cm.shape[1]):
        plt.text(
            j, i, f"{cm[i, j]}",
            horizontalalignment = "center",
            color = "white" if cm[i, j] > thresh else "black"
        )

plt.tight_layout()
plt.show()

_, axes = plt.subplots(nrows = 2, ncols = 5, figsize = (10, 5))
for ax, image, true_label, prediction in zip(axes.flatten(), X_test[:10], y_test[:10], predicted[:10]):
    image = image.reshape(28, 28)
    ax.set_axis_off()
    ax.imshow(image, cmap = plt.cm.gray_r, interpolation = "nearest")
    ax.set_title(f"True: {true_label}\nPred: {prediction}")

plt.show()
Start learning at 2024-12-03 12:03:13.908212 Stop learning at 2024-12-03 12:03:42.315214 Elapsed learning time: 0:00:28.407002 Classification Report: precision recall f1-score support 0 0.98 1.00 0.99 213 1 1.00 0.99 0.99 235 2 0.95 0.97 0.96 225 3 0.97 0.95 0.96 187 4 0.97 0.97 0.97 175 5 0.98 0.97 0.98 198 6 0.99 0.96 0.97 181 7 0.96 0.99 0.98 192 8 0.96 0.97 0.96 193 9 0.99 0.97 0.98 201 accuracy 0.98 2000 macro avg 0.98 0.97 0.98 2000 weighted avg 0.98 0.98 0.98 2000 Accuracy: 0.9755
Image in a Jupyter notebookImage in a Jupyter notebook
"""
Looped the Gamma Hyperparameter with C = 1
"""

import numpy as np
import matplotlib.pyplot as plt
from sklearn.datasets import fetch_openml
from sklearn.model_selection import train_test_split
from sklearn import svm, metrics
import datetime as dt

mnist = fetch_openml('mnist_784', version=1, as_frame = False)

X = mnist.data
y = mnist.target.astype(int)

random_indices = np.random.choice(len(X), 10000, replace = False)
X_subset = X[random_indices]
y_subset = y[random_indices]

X_subset = X_subset / 255.0

X_train, X_test, y_train, y_test = train_test_split(X_subset, y_subset, test_size = 0.2, random_state = 42)

param_C = 1.0
gamma_values = np.arange(0.001, 1, 0.005)

accuracies = []
learning_times = []

for gamma in gamma_values:
    classifier = svm.SVC(C = param_C, gamma = gamma)

    start_time = dt.datetime.now()
    classifier.fit(X_train, y_train)
    end_time = dt.datetime.now()

    elapsed_time = (end_time - start_time).total_seconds()
    learning_times.append(elapsed_time)

    predicted = classifier.predict(X_test)
    accuracy = metrics.accuracy_score(y_test, predicted)
    accuracies.append(accuracy)

    print(f"Gamma: {gamma:.5f}, Accuracy: {accuracy:.4f}, Learning Time: {elapsed_time:.2f}s")

"""
Values were taken out and a graph was made using matplotlib.
"""

"""
Looping the C Hyperparameter with Gamma = 0.026, the value of Gamma where accuracy peaked.
"""

import numpy as np
import matplotlib.pyplot as plt
from sklearn.datasets import fetch_openml
from sklearn.model_selection import train_test_split
from sklearn import svm, metrics
import datetime as dt

mnist = fetch_openml('mnist_784', version=1, as_frame = False)

X = mnist.data
y = mnist.target.astype(int)

random_indices = np.random.choice(len(X), 10000, replace = False)
X_subset = X[random_indices]
y_subset = y[random_indices]

X_subset = X_subset / 255.0

X_train, X_test, y_train, y_test = train_test_split(X_subset, y_subset, test_size = 0.2, random_state = 42)

param_C_values = np.arange(0.01, 2, 0.01)
gamma = 0.026

accuracies = []
learning_times = []

for param_C in param_C_values:
    classifier = svm.SVC(C = param_C, gamma = gamma)

    start_time = dt.datetime.now()
    classifier.fit(X_train, y_train)
    end_time = dt.datetime.now()

    elapsed_time = (end_time - start_time).total_seconds()
    learning_times.append(elapsed_time)

    predicted = classifier.predict(X_test)
    accuracy = metrics.accuracy_score(y_test, predicted)
    accuracies.append(accuracy)

    print(f"C Parameter: {param_C:.5f}, Accuracy: {accuracy:.4f}, Learning Time: {elapsed_time:.2f}s")

#Moving on to MLP with Quantum Approach

#Importing Necessary Libraries
!pip install pennylane
import pennylane as qml
import numpy as np
from sklearn.datasets import fetch_openml
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import cross_val_score
from sklearn.neural_network import MLPClassifier
Requirement already satisfied: pennylane in /usr/local/lib/python3.10/dist-packages (0.39.0) Requirement already satisfied: numpy<2.1 in /usr/local/lib/python3.10/dist-packages (from pennylane) (1.26.4) Requirement already satisfied: scipy in /usr/local/lib/python3.10/dist-packages (from pennylane) (1.13.1) Requirement already satisfied: networkx in /usr/local/lib/python3.10/dist-packages (from pennylane) (3.4.2) Requirement already satisfied: rustworkx>=0.14.0 in /usr/local/lib/python3.10/dist-packages (from pennylane) (0.15.1) Requirement already satisfied: autograd in /usr/local/lib/python3.10/dist-packages (from pennylane) (1.7.0) Requirement already satisfied: toml in /usr/local/lib/python3.10/dist-packages (from pennylane) (0.10.2) Requirement already satisfied: appdirs in /usr/local/lib/python3.10/dist-packages (from pennylane) (1.4.4) Requirement already satisfied: autoray>=0.6.11 in /usr/local/lib/python3.10/dist-packages (from pennylane) (0.7.0) Requirement already satisfied: cachetools in /usr/local/lib/python3.10/dist-packages (from pennylane) (5.5.0) Requirement already satisfied: pennylane-lightning>=0.39 in /usr/local/lib/python3.10/dist-packages (from pennylane) (0.39.0) Requirement already satisfied: requests in /usr/local/lib/python3.10/dist-packages (from pennylane) (2.32.3) Requirement already satisfied: typing-extensions in /usr/local/lib/python3.10/dist-packages (from pennylane) (4.12.2) Requirement already satisfied: packaging in /usr/local/lib/python3.10/dist-packages (from pennylane) (24.2) Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests->pennylane) (3.4.0) Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests->pennylane) (3.10) Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests->pennylane) (2.2.3) Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests->pennylane) (2024.8.30) 
#Loading MNIST Dataset
mnist = fetch_openml('mnist_784',version=1)
X,y = mnist.data / 255.0, mnist.target.astype(int)
X_train , X_test ,y_train , y_test = train_test_split(X,y,test_size=0.2,random_state=42)


#feature scaling and data preprocessing
scaler = StandardScaler()
X_train = scaler.fit_transform(X_train)
X_test = scaler.transform(X_test)
print(f"Training data dimension:{X_train.shape},Test data dimension:{X_test.shape}")
print(f"Training labels dimension:{y_train.shape},Test labels dimension:{y_test.shape}")
Training data dimension:(8000, 784),Test data dimension:(2000, 784) Training labels dimension:(8000,),Test labels dimension:(2000,) 
#Quantum Device setup
n_qubits = 3
dev = qml.device('default.qubit',wires=(n_qubits))

#Defining Quantum Circuits
@qml.qnode(dev)
def Quantum_Circ(p):
  for i in range(n_qubits):
    qml.RX(p[i],wires=i)
  return[qml.expval(qml.PauliZ(i)) for i in range(n_qubits)]

#Visualizing and Testing the Quantum Circuit
test_params = np.random.uniform(-np.pi,np.pi,n_qubits)
print('Quantum Circuit:')
print(qml.draw(Quantum_Circ)(test_params))
print(f"Testing quantum circuit with params:{test_params}")
qvalues = Quantum_Circ(test_params)
print(f"Quantum circuit output:{qvalues}")
Quantum Circuit: 0: ──RX(3.00)──┤ <Z> 1: ──RX(-1.10)─┤ <Z> 2: ──RX(-0.18)─┤ <Z> Testing quantum circuit with params:[ 2.99742779 -1.09737333 -0.17626929] Quantum circuit output:[-0.9896262315339532, 0.45593546069885144, 0.9845047525219398] 
#Translating Quantum Output to Hyperparameters
def translate_quantum_to_classic(qvalues):
  #learning_rate- Maps qvalues between -1 and 1 to Appropriate range between 0.0001 and 0.1
  learning_rate = 0.0001 + (0.1 - 0.0001)*((qvalues[0] + 1)/2)
  #num_neurons- Maps qvalues between -1 and 1 to Appropriate range between 16 and 256
  num_neurons = int(16 + (256 - 16)*((qvalues[1] + 1 / 2)))
  #alpha- Maps qvalue between -1 and 1 to Appropriate range between 0.0001 and 0.1
  alpha = 0.0001 + (0.1 - 0.0001)*((qvalues[2] + 1 / 2))
  return learning_rate,num_neurons,alpha

print(f"Translated hyperparameter:{translate_quantum_to_classic(qvalues)}")
Translated hyperparameter:(0.0006181697348790358, 245, 0.14840202477694178) 
#Fitness Function for Mlp Evalution
def fitness_function(qvalues):
  try:
    learning_rate , num_neurons , alpha = translate_quantum_to_classic(qvalues)
    model = MLPClassifier(hidden_layer_sizes=(num_neurons,),learning_rate_init=learning_rate,alpha = alpha, max_iter=20)
    cv_scores = cross_val_score(model,X_train,y_train,cv=3,scoring='accuracy')
    print(f"cross-validation scores:{cv_scores}")
    return np.mean(cv_scores)
  except Exception as e:
    print(f"Error during fitness evalution:{e}")
    return -np.pi

#Running the Fitness Function
test_qvalues = Quantum_Circ(test_params)
print(f"Fitness for test quantum values{test_qvalues}:{fitness_function(test_qvalues)}")
/usr/local/lib/python3.10/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:690: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (20) reached and the optimization hasn't converged yet. warnings.warn( /usr/local/lib/python3.10/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:690: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (20) reached and the optimization hasn't converged yet. warnings.warn(
cross-validation scores:[0.94375703 0.9343832 0.94073518] Fitness for test quantum values[-0.9896262315339532, 0.45593546069885144, 0.9845047525219398]:0.9396251387556301
/usr/local/lib/python3.10/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:690: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (20) reached and the optimization hasn't converged yet. warnings.warn(

Evaluation Results :

Cross-Validation scores: Fold1: 0.94375703 ,Fold2: 0.9343832 , Fold3: 0.94073518

Average Accuracy: 0.9396251387556301

Fitness for test quantum values: [-0.9896262315339532, 0.45593546069885144, 0.9845047525219398]

#Visualizing Cross-Validation fold scores and Fitness score For Model  Performance Comparison using Bar Graph
import matplotlib.pyplot as plt
import numpy as np
cv_scores = [0.94375703,0.9343832,0.94073518]
fitness_score = 0.9396251387556301
labels = ['Fold 1', 'Fold 2', 'Fold 3']
plt.figure(figsize=(8,6))
bars = plt.bar(labels, cv_scores, color=['#4CAF50','#2196F3','#FFC107'],alpha=0.8)
plt.axhline(y=fitness_score, color='red', linestyle='--',linewidth=1.5,label=f'Fitness score:{fitness_score:4f}')
for bar,score in zip(bars, cv_scores):
  plt.text(bar.get_x() + bar.get_width() / 2, bar.get_height() + 0.003,f'{score:.4f}', ha='center',fontsize=10)
plt.title('Cross-Validation scores and Fitness score', fontsize=14)
plt.ylabel('Accuracy', fontsize=12)
plt.xlabel('Cross-Validation Folds', fontsize=12)
plt.legend()
plt.tight_layout()
plt.show()
Image in a Jupyter notebook