ishwor_basic_keras_model_on_iris_dataset.py

# Thanks to Jose Portilla on Udemy for giving me amazing idea for this model

# Performing keras basics based on Iris dataset

# First of all, importing basic libraries, numpy and pandas
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import keras

# Importing the Iris dataset
from sklearn.datasets import load_iris

# Initializing dataset
iris = load_iris()

# Checking the type of dataset
type(iris)

# checking the description of the dataset
print(iris.DESCR)

# Grabbing features and labels
X = iris.data
y = iris.target

from keras.utils import to_categorical
y = to_categorical(y)


y.shape

# Importing train_test_split module
from sklearn.model_selection import train_test_split
# Splitting the dataset into training and test set with the help of train_test_split module of sklearn.
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42)

# Importing MinMaxScaler from scikit-learn and initiating it
from sklearn.preprocessing import MinMaxScaler
scaler_object = MinMaxScaler()

# Fitting the scaler object to training data to scale the dataset into same level
scaler_object.fit(X_train)

# splitting the train and test dataset after scaling and transforming
scaled_X_train = scaler_object.transform(X_train)
scaled_X_test = scaler_object.transform(X_test)

# Importing sequential and dense from keras
from keras.models import Sequential
from keras.layers import Dense

model = Sequential()
model.add(Dense(8,input_dim=4, activation='relu'))
model.add(Dense(8,input_dim=4, activation='relu'))
model.add(Dense(3, activation='softmax'))
model.compile(loss = 'categorical_crossentropy', optimizer='adam', metrics = ['accuracy'])

# Checking the summary of the model
model.summary()

# Fitting the model with 150 epochs
model.fit(scaled_X_train, y_train, epochs=150, verbose=2)

# Need to scale new incoming data
model.predict_classes(scaled_X_test)
predictions = model.predict_classes(scaled_X_test)

y_test.argmax(axis=1)

# Importing confusion matrix, classification report and the accuracy score module from keras
from sklearn.metrics import confusion_matrix, classification_report, accuracy_score
# Printing confusion matrix
confusion_matrix(y_test.argmax(axis=1), predictions)
# Printing classification report
print(classification_report(y_test.argmax(axis=1), predictions))
# accuracy score
accuracy_score(y_test.argmax(axis=1), predictions)

# Saving the model in the format of h5 file in order to run in the future of deploy it
model.save('ishwor_basic_keras_model_on_Iris_Dataset.h5')

# If I want to run the model again
my_new_model = load_model('ishwor_basic_keras_model_on_Iris_Dataset.h5')

# If I want to make some predictions into trained model above, let's say
model.predict_classes(scaled_X_test)