knn assignments in machine learning

Assignment 1 — Employee Salary Category Prediction using K-Nearest Neighbors (KNN)

Objective

Build a K-Nearest Neighbors (KNN) Classification model to predict whether an employee belongs to the High Salary or Low Salary category.

Dataset

File Name

employee_salary_knn.csv

EmployeeID	ExperienceYears	Education	Skills	InterviewScore	SalaryCategory
1	1	Bachelors	Python	55	Low
2	2	Bachelors	Java	58	Low
3	3	Masters	Python	65	Low
4	4	Masters	Java	68	Low
5	5	Masters	Python	72	High
6	6	Masters	Python	75	High
7	7	PhD	Python	80	High
8	8	PhD	Data Science	85	High
9	2	Diploma	Java	52	Low
10	3	Diploma	Java	60	Low
11	4	Bachelors	Python	66	Low
12	5	Bachelors	Data Science	70	High
13	6	Masters	Data Science	74	High
14	7	Masters	Python	79	High
15	8	PhD	AI	90	High
16	1	Diploma	Java	50	Low
17	2	Diploma	Python	57	Low
18	3	Bachelors	Java	63	Low
19	4	Masters	AI	69	Low
20	5	Masters	AI	73	High
21	6	Masters	Python	76	High
22	7	PhD	AI	82	High
23	8	PhD	Data Science	88	High
24	2	Diploma	Python	56	Low
25	3	Bachelors	Python	64	Low
26	4	Bachelors	Java	67	Low
27	5	Masters	Data Science	71	High
28	6	Masters	AI	77	High
29	7	PhD	Python	83	High
30	8	PhD	AI	91	High
31	1	Diploma	Java	51	Low
32	2	Bachelors	Java	59	Low
33	3	Bachelors	Python	62	Low
34	4	Masters	Data Science	68	Low
35	5	Masters	Python	72	High
36	6	Masters	AI	75	High
37	7	PhD	AI	81	High
38	8	PhD	Data Science	86	High
39	2	Diploma	Python	54	Low
40	3	Diploma	Java	61	Low
41	4	Bachelors	AI	66	Low
42	5	Bachelors	Data Science	70	High
43	6	Masters	Python	74	High
44	7	Masters	AI	80	High
45	8	PhD	Python	87	High
46	2	Diploma	Java	53	Low
47	3	Bachelors	Python	65	Low
48	5	Masters	AI	73	High
49	6	Masters	Data Science	78	High
50	8	PhD	AI	92	High

---

Before Training

This dataset cannot be used directly.

The following preprocessing is required.

Step 1 — Handle Missing Values (if any)

Check whether the dataset contains missing values.

---

Step 2 — Encode Categorical Columns

Use Label Encoding for:

• Education

• Skills

• SalaryCategory

Example

Original	Encoded
Diploma	0
Bachelors	1
Masters	2
PhD	3

---

Step 3 — Feature Scaling

Since KNN is a distance-based algorithm, apply StandardScaler (or MinMaxScaler) to the input features before training.

---

Features (X)

ExperienceYears
Education
Skills
InterviewScore

---

Target (y)

SalaryCategory

---

Practical Questions

Question 1

Load the dataset using Pandas.

---

Question 2

Check the first five records.

---

Question 3

Display the shape of the dataset.

---

Question 4

Check the data types of all columns.

---

Question 5

Check whether the dataset contains missing values.

---

Question 6

Apply Label Encoding to all categorical columns.

---

Question 7

Separate the features (X) and target (y).

---

Question 8

Split the dataset into training and testing sets using an 80:20 ratio.

---

Question 9

Apply StandardScaler to the training and testing data.

---

Question 10

Create a KNN classifier with:

K = 3

Train the model.

---

Question 11

Predict the salary category for the test dataset.

---

Question 12

Calculate the Accuracy of the model.

---

Question 13

Generate the Confusion Matrix.

---

Question 14

Calculate:

• Precision

• Recall

• F1-Score

---

Question 15

Repeat the experiment using:

• K = 5

• K = 7

• K = 9

Compare the accuracy of each model.

---

Question 16

Which value of K gives the best accuracy?

---

Question 17

Predict the salary category of the following employee:

Experience	Education	Skills	Interview Score
6 Years	Masters	AI	78

# ==========================================

# Employee Salary Category Prediction using KNN

# ==========================================

import pandas as pd

# Machine Learning Libraries

from sklearn.model_selection import train_test_split

from sklearn.preprocessing import LabelEncoder

from sklearn.preprocessing import StandardScaler

from sklearn.neighbors import KNeighborsClassifier

# Evaluation Metrics

from sklearn.metrics import accuracy_score

from sklearn.metrics import confusion_matrix

from sklearn.metrics import precision_score

from sklearn.metrics import recall_score

from sklearn.metrics import f1_score

from sklearn.metrics import classification_report

# ==========================================

# Load Dataset

# ==========================================

df = pd.read_csv("employee_salary_knn.csv")

print("Dataset Loaded Successfully\n")

# ==========================================

# Display Dataset

# ==========================================

print("First 5 Records")

print(df.head())

print("\nLast 5 Records")

print(df.tail())

print("\nShape")

print(df.shape)

print("\nInformation")

print(df.info())

print("\nData Types")

print(df.dtypes)

print("\nMissing Values")

print(df.isnull().sum())

# ==========================================

# Label Encoding

# ==========================================

education_encoder = LabelEncoder()

skills_encoder = LabelEncoder()

salary_encoder = LabelEncoder()

df["Education"] = education_encoder.fit_transform(df["Education"])

df["Skills"] = skills_encoder.fit_transform(df["Skills"])

df["SalaryCategory"] = salary_encoder.fit_transform(df["SalaryCategory"])

print("\nEncoded Dataset")

print(df.head())

# ==========================================

# Features and Target

# ==========================================

X = df[

    [

        "ExperienceYears",

        "Education",

        "Skills",

        "InterviewScore"

    ]

]

y = df["SalaryCategory"]

# ==========================================

# Split Dataset

# ==========================================

X_train, X_test, y_train, y_test = train_test_split(

    X,

    y,

    test_size=0.20,

    random_state=42

)

# ==========================================

# Feature Scaling

# ==========================================

scaler = StandardScaler()

X_train = scaler.fit_transform(X_train)

X_test = scaler.transform(X_test)

# ==========================================

# Create KNN Model

# ==========================================

model = KNeighborsClassifier(

    n_neighbors=3

)

# ==========================================

# Train Model

# ==========================================

model.fit(

    X_train,

    y_train

)

print("\nModel Trained Successfully")

# ==========================================

# Prediction

# ==========================================

y_pred = model.predict(X_test)

print("\nActual Values")

print(y_test.values)

print("\nPredicted Values")

print(y_pred)

# ==========================================

# Accuracy

# ==========================================

accuracy = accuracy_score(

    y_test,

    y_pred

)

print("\nAccuracy")

print(accuracy)

print("Accuracy Percentage :", accuracy * 100)

# ==========================================

# Confusion Matrix

# ==========================================

cm = confusion_matrix(

    y_test,

    y_pred

)

print("\nConfusion Matrix")

print(cm)

# ==========================================

# Precision

# ==========================================

precision = precision_score(

    y_test,

    y_pred

)

print("\nPrecision")

print(precision)

# ==========================================

# Recall

# ==========================================

recall = recall_score(

    y_test,

    y_pred

)

print("\nRecall")

print(recall)

# ==========================================

# F1 Score

# ==========================================

f1 = f1_score(

    y_test,

    y_pred

)

print("\nF1 Score")

print(f1)

# ==========================================

# Classification Report

# ==========================================

print("\nClassification Report")

print(

    classification_report(

        y_test,

        y_pred

    )

)

# ==========================================

# Compare Different K Values

# ==========================================

print("\n==============================")

print("Compare Different K Values")

print("==============================")

for k in [3, 5, 7, 9]:

    model = KNeighborsClassifier(

        n_neighbors=k

    )

    model.fit(

        X_train,

        y_train

    )

    prediction = model.predict(X_test)

    accuracy = accuracy_score(

        y_test,

        prediction

    )

    print("K =", k, "Accuracy =", accuracy)

# ==========================================

# Best K

# ==========================================

best_accuracy = 0

best_k = 0

for k in range(1, 16):

    model = KNeighborsClassifier(

        n_neighbors=k

    )

    model.fit(

        X_train,

        y_train

    )

    prediction = model.predict(X_test)

    accuracy = accuracy_score(

        y_test,

        prediction

    )

    if accuracy > best_accuracy:

        best_accuracy = accuracy

        best_k = k

print("\nBest K :", best_k)

print("Best Accuracy :", best_accuracy)

# ==========================================

# Predict New Employee

# ==========================================

new_employee = pd.DataFrame(

    {

        "ExperienceYears": [6],

        "Education": ["Masters"],

        "Skills": ["AI"],

        "InterviewScore": [78]

    }

)

# Encode using previously trained encoders

new_employee["Education"] = education_encoder.transform(

    new_employee["Education"]

)

new_employee["Skills"] = skills_encoder.transform(

    new_employee["Skills"]

)

# Scale

new_employee = scaler.transform(new_employee)

# Prediction

prediction = model.predict(new_employee)

print("\nPrediction")

if prediction[0] == 1:

    print("High Salary Category")

else:

    print("Low Salary Category")

Assignment 2 — Loan Approval Prediction using K-Nearest Neighbors (KNN)

Objective

Build a K-Nearest Neighbors (KNN) model to predict whether a customer's loan application will be Approved or Rejected.

Banks often compare a new applicant with previous applicants having similar income, employment, credit history, and education. This makes KNN a suitable algorithm for this problem.

---

Dataset

File Name

loan_approval_knn.csv

CustomerID	Age	Education	EmploymentType	MonthlyIncome	CreditScore	ExistingLoans	City	LoanApproved
1	24	Graduate	Private	25000	620	0	Delhi	No
2	26	Graduate	Private	28000	640	1	Delhi	No
3	28	Post Graduate	Government	35000	690	1	Mumbai	Yes
4	30	Post Graduate	Government	42000	730	1	Mumbai	Yes
5	22	Graduate	Private	22000	610	0	Jaipur	No
6	32	Post Graduate	Government	50000	760	2	Delhi	Yes
7	35	PhD	Government	65000	810	1	Bangalore	Yes
8	27	Graduate	Private	30000	660	1	Pune	No
9	29	Graduate	Business	45000	710	1	Delhi	Yes
10	31	Post Graduate	Business	52000	750	2	Mumbai	Yes
11	25	Graduate	Private	27000	630	0	Delhi	No
12	33	PhD	Government	70000	820	1	Bangalore	Yes
13	34	Post Graduate	Business	56000	770	2	Mumbai	Yes
14	23	Graduate	Private	24000	615	0	Jaipur	No
15	36	PhD	Government	72000	830	1	Delhi	Yes
16	26	Graduate	Private	29000	645	1	Pune	No
17	28	Graduate	Business	40000	700	1	Delhi	Yes
18	30	Post Graduate	Government	48000	740	2	Mumbai	Yes
19	21	Graduate	Private	21000	600	0	Jaipur	No
20	37	PhD	Government	76000	840	2	Bangalore	Yes
21	24	Graduate	Private	26000	625	0	Delhi	No
22	27	Graduate	Business	39000	695	1	Pune	Yes
23	29	Post Graduate	Government	46000	725	1	Mumbai	Yes
24	31	Business	Business	54000	760	2	Delhi	Yes
25	22	Graduate	Private	23000	610	0	Jaipur	No
26	34	PhD	Government	68000	815	1	Bangalore	Yes
27	28	Graduate	Private	31000	665	1	Delhi	No
28	30	Business	Business	50000	745	2	Mumbai	Yes
29	35	Post Graduate	Government	62000	800	1	Delhi	Yes
30	23	Graduate	Private	24000	620	0	Jaipur	No
31	26	Graduate	Private	28000	640	1	Delhi	No
32	32	Government	Government	53000	755	2	Mumbai	Yes
33	36	PhD	Government	74000	835	1	Bangalore	Yes
34	24	Graduate	Private	25500	628	0	Delhi	No
35	27	Graduate	Business	41000	705	1	Pune	Yes
36	29	Post Graduate	Government	47000	730	2	Mumbai	Yes
37	22	Graduate	Private	22500	608	0	Jaipur	No
38	34	Business	Business	59000	785	1	Delhi	Yes
39	31	Post Graduate	Government	51000	750	2	Mumbai	Yes
40	37	PhD	Government	78000	845	2	Bangalore	Yes
41	25	Graduate	Private	27500	635	0	Delhi	No
42	28	Graduate	Business	43000	715	1	Pune	Yes
43	30	Post Graduate	Government	49500	740	2	Mumbai	Yes
44	21	Graduate	Private	21500	602	0	Jaipur	No
45	35	PhD	Government	71000	825	1	Delhi	Yes
46	27	Graduate	Private	30000	660	1	Delhi	No
47	33	Business	Business	57000	775	2	Mumbai	Yes
48	29	Post Graduate	Government	48500	735	2	Delhi	Yes
49	23	Graduate	Private	23500	618	0	Jaipur	No
50	38	PhD	Government	80000	850	2	Bangalore	Yes

---

Data Cleaning Required

Check the dataset for:

• Missing values

• Duplicate records

• Incorrect data types

---

Encoding Required

Use Label Encoding for the following columns:

• Education

• EmploymentType

• City

• LoanApproved

---

Feature Scaling Required

Since KNN is a distance-based algorithm, apply:

• StandardScaler (recommended)

or

• MinMaxScaler

---

Features (X)

Age
Education
EmploymentType
MonthlyIncome
CreditScore
ExistingLoans
City

---

Target (y)

LoanApproved

---

Practical Questions

Question 1

Load the dataset using Pandas.

Question 2

Display the first five records.

Question 3

Display the last five records.

Question 4

Display the shape of the dataset.

Question 5

Display the information (info()) of the dataset.

Question 6

Check all missing values.

Question 7

Check duplicate records.

Question 8

Apply Label Encoding to all categorical columns.

Question 9

Separate the input features (X) and target (y).

Question 10

Split the dataset into training and testing sets (80:20).

Question 11

Apply StandardScaler to the training and testing data.

Question 12

Train a KNN Classifier with:

K = 5

Question 13

Predict the loan approval status of the test dataset.

Question 14

Calculate the model accuracy.

Question 15

Generate the confusion matrix.

Question 16

Calculate:

• Precision

• Recall

• F1-Score

Question 17

Display the classification report.

Question 18

Compare the model performance for:

• K = 3

• K = 5

• K = 7

• K = 9

• K = 11

Which value of K gives the highest accuracy?

Question 19

Predict the loan approval status of the following new customer:

Age	Education	EmploymentType	MonthlyIncome	CreditScore	ExistingLoans	City
30	Post Graduate	Government	52000	760	1	Delhi