grade5.py

# IMPORTING LIBRARIES ***********************************************************
import pandas as pd
import matplotlib.pyplot as plt 
import seaborn as sns
sns.set_palette('husl')

from sklearn.model_selection import train_test_split
from sklearn.model_selection import cross_val_score
from sklearn.model_selection import StratifiedKFold
from sklearn.metrics import classification_report
from sklearn.metrics import accuracy_score

from sklearn.svm import SVC

# Load The Training and Test Data
train_df = pd.read_csv('student_marks_train.csv')
result = train_df.info()
print("Details of the Train DataFrame:")
print(result)

test_df = pd.read_csv('student_marks_test.csv')
result = test_df.info()
print("Details of the Test DataFrame:")
print(result)

# EXPLORE the data to see the distributions ***************************************.
sns.pairplot(train_df, hue='class-label', markers='+')
plt.show()

# PRE-PROCESS the data to see if you can improve the final classification**********


# SPLIT taining data to test ML model *********************************************
# Drop the column that contains the class labels
X = train_df.drop(['class-label'], axis=1)
y = train_df['class-label']
print(f'X shape: {X.shape} | y shape: {y.shape} ')

# Split: 50% for training. 50% for testing
# We have to split the given training data to test since the test data given does not have class-labels
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(
                                                    X, 
                                                    y, 
                                                    test_size=0.50, 
                                                    random_state=1)
            
# BUILD and TRAIN ML model *******************************************************
# Support vector machine algorithm for ML
from sklearn.svm import SVC
svn = SVC()
svn.fit(X_train, y_train)
# Model evaluation
# Check predictions of the test dataset that we created
predictions = svn.predict(X_test)
# Calculate the accuracy
from sklearn.metrics import accuracy_score
print("Current Accuracy =", accuracy_score(y_test, predictions))

# TEST the class labels for the given Test set ***********************************
# Now Predict the given test set using the ML model that we created
test_labels = (svn.predict(test_df))
print("------CLASSIFICATION RESLUTS START------")
for i in range(len(test_labels)):
  print(test_labels[i])
print("------CLASSIFICATION RESLUTS END------")

# You can cut and paste the output to the answer box and click Check to see 
# how good your ML model is working on the new testset where you do not know the class-lables
# Try to see if you can get it 100% (1.0)