LAB EXERCISE - 14 Naive Bayes lassifier 1. Aim of the Experiment: Implement and demonstrate the working of Naive Bayesian classifier using a sample data set. Build the model to classify a test sample. 2. Reference to Text book for Algorithms: Refer to Section 8.3.1 in Chapter 8 Bayesian Learning to understand the working of the algorithm, Listing 1: Sample Dataset Used: Table 8.1 Table 8.1 Trai 2 [28 [No Good “Moderate Yes 3. [29 [No ‘Average Poor ‘No 4 [8 [No ‘Average Good ‘No 5. [58 Yes Good Moderate Yes 6 |e ‘Yes Good Moderate Yes 7 [e8 [Yes Good Poor ‘No 8 [29 [No Very good | Good Yes 9 [B8 [Yes Good Good Yes 10. [28 [Yes ‘Average Good Yes 3. Python Program with Explanation: 1, Import LabelEncoder to normalize labels. from sklearn.preprocessing import LabelEncoder 2. Import train_test_split function, Copyright @ Oxford University Press, India 2021from sklearn.model_selection import train_test_spli 3. Import Gaussian Classifier from sklearn.naive_bayes. from sklearn.naive_bayes import GaussianNB 4, Import preprocessing package to use transformer classes. from sklearn import preprocessing 5. Import classification_report and confusion_matrix from. sklearn.mettics to measure the quality of predictions, from sklearn.metrics import classification_report, confusion_matrix 6. Create lists, CGPA, Inter, PK, CS and Job. CGPA = ['g9','g8','29','18','g8'/'g9',18''e9''e8'/e8'] Inter = ['Y','N','N4'N' "YY “Y PK = [4H Fat 254 4444'S] CS =['G)'M''P'6)'M'M!/'P)'G''G''G'] Job = [V4.4 /NY'NI YY YUINE SEY V5] 7. Create labelEncoder le to encode labels with value between 0 and no_of_classes-1 le = preprocessing.LabelEncoder() 8. Convert non-numeric labels of all features into numbers. Then print and sce the encoded labels. CGPA_encoded = le fit_transform(CGPA) print("CGPA:", CGPA_encoded) Inter_encoded = le-fit_transform(Inter) PK_encoded = le.fit_transform(PK) CS_encoded = le.fit_transform(CS) label = le.fit_transform(Job) print("Inter:",Inter_encoded) print ("PK:",PK_encoded) Copyright @ Oxford University Press, India 2021print ("CS:",CS_encoded) print("Job:" label) 9. Create a list/dynamic array called “features features =] 10. Append all the encoded features to the list created. for iin range(len(CGPA_encoded)): features.append{[CGPA_encoded!i, Inter_encodedfi], PK_encodedfi, 5_encoded{ill) 11, Split the dataset into training dataset and test dataset by using the function train_test_splitQ, X train,X_testyy_train,y_test=train_test_split(features,label,test_size=0.30,random_s tate=2) 12, Create a Gaussian Classifier, model = GaussianNB() 13, Train the model using the training sets. model fit(features, label) 14, After training, use the fitted model to predict a new instance. y_pred = model.predict(X_test) 15. Generate classification report & confusion matrix to measure the quality of predictions. print(classification_report(y_test, y_pred)) print(confusion_matrixty_test, y_pred)) 16. Predict Output. Copyright @ Oxford University Press, India 2021# The non-numerical equivalent of the new instance [2, 0, 2, 0] given is [2:'18", 0:°N 2: “G'] print({2,0,2,0]) if model,predict{{{2,0,2,0]) print("Predicted Value:Got JOB",predicted ) else: print("Didnt get JOB") # The non-numerical equivalent of the new instance [0, 1, 0, 1] given is (0:"g8 1,08", 1M] print({0,1,0,1]) if model.predict({{0,1,0,11) print(""Predicted Value:Got JOB" predicted ) print("Didnt get 108") Complete Program: from sklearn preprocessing import LabelEncoder from sklearn.model_selection import train_test_split from sklearn.nalve_bayes import GaussianNB {rom sklearn import preprocessing from sklearn.metrics import classification_report, confusion_mateix CPA = [9 Inter = [¥ creating labelEncoder reprocessing.LabelEncoder() +# Converting string labels into numbers. Copyright @ Oxford University Press, India 2021left transform(coPa) print("CGPA:", CGPA_encoded) Inter_encoded = e.it_transform(inter) PK_encoded = lesit_transform(PK) CS_encoded = le.fit_transform(CS) label le.ft_transformUlob) print("Inter:",Inter_encoded)} print ("PK:",PK_encoded) print ("CS:",CS_encoded} print("Job:" label) features = (} for iin range(len(CGPA_encoded)}: features.append([CGPA_encodedlil, Inter_encoded[i], PK_encoded{i], CS_encoded{il}) X train,X_testy_train,y_test=train_test_split{features,label,test_size=0.30,random_state=2) Create a Gaussian Classifier ‘model = GaussianNB() #4 Train the model using the training sets model it(features,label) Predict Output y_pred = model.predict(X_test) print(classification_reportly_test, y_pred)) print(confusion_matrix(y_test, y_predl) print([2,0,2,0}) if model predict({l2,0,2,0) print("Predicted Value:Got JOB") print("Predicted Value:Didn't get JOB") Copyright @ Oxford University Press, India 2021print({0,2,0,2]) if model.predict({(0,1,0,1) print("Predicted Value:Got JOB") print("Predicted Value:Didn't get JO Outpu Python 3.8.3 (tags/v3.8.3:6f8¢832, May 13 2020, 22:37:02) [MSC v.1924 64 bit (AMDS4)] on win32 ‘Type "help", "copyright", “credits” or “license()" for more information. RESTART: C:\Users\ADMIN\pythonpgms\inal\jnf naive bayes sklearn test. py capa: (1012012100) Inter: [1000111011] Pk: [1022000102] ¢5:[0120112000) Job: [1100110113] precision recall f1-score support 1 100 100 100 3 accuracy 1003 macroavg 1.00 1.00 100 3 weighted avg 1.00 1.00 100 3 (0) (2,0, 2,0] Predicted Value:Didn't get JOB (0,1, 0,1) Predicted Value:Got JOB Copyright @ Oxford University Press, India 2021Screenshot of the Output: Listing 2: Program Code: from sklearn import datasets from sklearn import metrics from sklearn.naive_bayes import GaussianNB from sklearn.metrics import classification_report, confusion_matrix from skiearn.model_selection import train_test_split, ff Load the Iris dataset dataset = datasets.load_iris() 4 Fita Naive Bayes model to the data model = GaussianNa() X_train,X_testy_train,y_test=train_test_split(dataset data dataset target,test_s e=2) 0.30,random_stat model fit(X_train, y_train) Copyright @ Oxford University Press, India 2021print(model) 4## Make predictions vy.expected = _test y_predicted = model.predict(X_test) ## Evaluate the model and print the classification report, Confusion Matrix print(metrics.classification_report(y_expected, y_predicted)) print(metrics.confusion_matrix(y_expected, y_predicted)) Sereen Shot of the Progra Output: Python 3.8.3 (tags/v3.8.3:6f8cB32, May 13 2020, 22:37:02) [MSC v.1924 64 bit (AMD64)] on win32 ‘Type "help", "copyright", "credits" or "license()" for more information. RESTART: C:\Users\ADMIN\pythonpgms\Review inf naive bayes iris.py Copyright @ Oxford University Press, India 2021GaussianNB() precision recall fl-score support 0 100 100 100 17 1 100 093 097 15 2 093 100 096 13 accuracy 0.9845 macroavg 0.98 0.98 0.98 45 weighted avg 0.98 0.98 0.98 45 (127 0 9} [014 4) [0 013) Screen Shot of the Output: Copyright @ Oxford University Press, India 2021