ashu1069

import numpy as np
import matplotlib.pyplot as plt

from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split

from sklearn.metrics import classification_report
from sklearn.neighbors import KNeighborsClassifier

#load the iris dataset

ashu1069

def PCA(x, N):
    '''
    Arguments:
    x: input features
    N: number fo principal components
    '''
    #mean
    x_mean = np.mean(x)
    
    #standard deviation

ashu1069

def markov_chain(transition_matrix):
    # Create a directed graph.
    G = nx.DiGraph()

    # Nodes represent pages. Assume node labels are 0, 1, 2, ... for simplicity.
    num_nodes = transition_matrix.shape[0]
    G.add_nodes_from(range(num_nodes))

    # Iterate through the transition matrix to create edges.
    for i in range(num_nodes):

ashu1069

def PageRank(transition_matrix, d, max_iterations, conv_thres):
    '''
    Arguments:
    transition_matrix: a matrix or numpy array representing the probabilities of going from one page to another
    d: damping factor
    max_iterations: number of iterations
    conv_thres: convergence threshold
    
    Return: ranks of each webpage, as columns of the transition matrix
    '''

ashu1069

#Input image of size = 448*448
from keras.preprocessing.image import load_img, img_to_array
img = load_img("test2.png")

#Converting image to array
img = img_to_array(img)
img = np.expand_dims(img, axis=0)
#Predicting using the truncated model
feature_output = model_short.predict(img)

ashu1069

#Analyzing the filter outputs
#For that, we will define a truncated architecture to study the filter outputs
conv_layer_index=[0, 2, 4, 5, 6, 7, 9, 10, 11, 12]  #We can choose any number of convolution layers; simply find the indices of them from the layer array
outputs = [model.layers[i].output for i in conv_layer_index]
model_short = Model(inputs=model.inputs, outputs=outputs)
#The following command will plot the complete analysis of our architecture
print(model_short.summary()) 

ashu1069

#Studying and analyzing the layers of the YOLO architecture
layer = model.layers
filters, biases = model.layers[0].get_weights()    #the indices of layers can be chosen once you see the summary of the model
print(layer[0].name, filters.shape)

#Plotting the filters
fig1 = plt.figure(figsize=(8,8))
columns=8
rows=8
n_filters=columns*rows

ashu1069

#Rectified Layer Unit activation function and coefficient of the Leaky ReLU equation is 0.1
lrelu = keras.layers.LeakyReLU(alpha=0.1)

#Building the YOLO architecture
model = Sequential()
model.add(Conv2D(filters=64, kernel_size= (7, 7), strides=(1, 1), input_shape =(448, 448, 3), padding = 'same', activation=lrelu, kernel_regularizer=l2(5e-4)))
model.add(MaxPooling2D(pool_size=(2, 2), strides=(2, 2), padding = 'same'))

model.add(Conv2D(filters=192, kernel_size= (3, 3), padding = 'same', activation=lrelu, kernel_regularizer=l2(5e-4)))
model.add(MaxPooling2D(pool_size=(2, 2), strides=(2, 2), padding = 'same'))

ashu1069

import matplotlib.pyplot as plt
import numpy as np

from tensorflow import keras
from keras.models import Model
from keras.models import Sequential
from keras.layers import Dense, Dropout, Flatten
from keras.layers import Conv2D, MaxPooling2D
from keras.regularizers import l2