Scribd
Upload
8 views13 pages

USA Housing Dataset Machine Learning Analysis

The document outlines an assignment on Machine Learning using the USA Housing dataset, focusing on housing prices and various features. It details the steps for data cleaning, normalization, and the implementation of Linear and Polynomial Regression models, including performance evaluation metrics. The results indicate that while Polynomial Regression provides similar performance to Linear Regression, it does not significantly improve generalization.

Uploaded by

kishlaychauhan777
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
8 views13 pages

USA Housing Dataset Machine Learning Analysis

The document outlines an assignment on Machine Learning using the USA Housing dataset, focusing on housing prices and various features. It details the steps for data cleaning, normalization, and the implementation of Linear and Polynomial Regression models, including performance evaluation metrics. The results indicate that while Polynomial Regression provides similar performance to Linear Regression, it does not significantly improve generalization.

Uploaded by

kishlaychauhan777
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Assignment

September 4, 2025

Anmol Shakya

Assignment 1 Machine Learning

2301330100037

Question 1

We are using the USA Housing dataset which contains information about housing prices along

with features such as average income, house age, number of rooms, number of bedrooms, population,
etc.

[4]:

0 79545.458574 5.682861 7.009188


1 79248.642455 6.002900 6.730821
2 61287.067179 5.865890 8.512727
3 63345.240046 7.188236 5.586729
4 59982.197226 5.040555 7.839388

Avg. Area Number of Bedrooms Area PopulaBon Price \


0 4.09 23086.800503 1.059034e+06
1 3.09 40173.072174 1.505891e+06
2 5.13 36882.159400 1.058988e+06
3 3.26 34310.242831 1.260617e+06
4 4.23 26354.109472 6.309435e+05

1
Address
0
208 Michael Ferry Apt. 674\nLaurabury, NE 3701…
1
188 Johnson Views Suite 079\nLake Kathleen, CA…
2
9127 Elizabeth Stravenue\nDanieltown, WI 06482…
3
USS Barne]\nFPO AP 44820
4
USNS Raymond\nFPO AE 09386

Question 2

Show raw data summary before and after cleaning.

Print key statistics (.describe(), .info() etc.).

Plot missing data heatmap before cleaning and after cleaning (use [Link] or similar).

[5] : # Info and statistics before cleaning


print("Dataset Info:") print([Link]())

print("\nDataset DescripBon:")
print([Link]())

# Checking missing values


print("\nMissing Values:")
print([Link]().sum())

Dataset Info:
<class '[Link]'>
RangeIndex: 5000 entries, 0 to 4999 Data
columns (total 7 columns):
# Column Non-Null Count Dtype

0 Avg. Area Income 5000 non-null float64


1 Avg. Area House Age 5000 non-null float64
2 Avg. Area Number of Rooms 5000 non-null float64
3 Avg. Area Number of Bedrooms 5000 non-null float64
4 Area PopulaBon 5000 non-null float64
5 Price 5000 non-null float64
6 Address 5000 non-null object dtypes: float64(6), object(1) memory usage: 273.6+ KB

2
None
Dataset DescripBon:

Avg. Area Income Avg. Area House Age Avg. Area Number of Rooms \

count 5000.000000 5000.000000 5000.000000


mean 68583.108984 5.977222 6.987792
std 10657.991214 0.991456 1.005833
min 17796.631190 2.644304 3.236194
25% 61480.562388 5.322283 6.299250
50% 68804.286404 5.970429 7.002902
75% 75783.338666 6.650808 7.665871
max 107701.748378 9.519088 10.759588

Avg. Area Number of Bedrooms Area PopulaBon Price


count 5000.000000 5000.000000 5.000000e+03
mean 3.981330 36163.516039 1.232073e+06
std 1.234137 9925.650114 3.531176e+05
min 2.000000 172.610686 1.593866e+04
25% 3.140000 29403.928702 9.975771e+05
50% 4.050000 36199.406689 1.232669e+06
75% 4.490000 42861.290769 1.471210e+06
max 6.500000 69621.713378 2.469066e+06

Missing Values:

Avg. Area Income 0

Avg. Area House Age 0

Avg. Area Number of Rooms 0

Avg. Area Number of Bedrooms 0

Area PopulaBon 0

Price 0

3
Handle missing values (drop/impute with justification).

Encode categorical features using appropriate techniques (Label Encoding / One-Hot Encoding).

Normalize/Standardize numerical columns if necessary.

Split your dataset into train and test (80:20 or 70:30).

##As we have checked there is no missing value USA housing dataset.

4
[7]: Avg. Area Income Avg. Area House Age Avg. Area Number of Rooms \
0 79545.458574 5.682861 7.009188
1 79248.642455 6.002900 6.730821
2 61287.067179 5.865890 8.512727
3 63345.240046 7.188236 5.586729
4 59982.197226 5.040555 7.839388

Avg. Area Number of Bedrooms Area PopulaBon Price


0 4.09 23086.800503 1.059034e+06
1 3.09 40173.072174 1.505891e+06
2 5.13 36882.159400 1.058988e+06
3 3.26 34310.242831 1.260617e+06
4 4.23 26354.109472 6.309435e+05

0.0.1 Normalization / Standardization

Since the dataset features (e.g., Income, Rooms, Population) are on very different scales, we
applied Standardization using StandardScaler.
This scales all numerical features to have mean = 0 and standard deviation = 1, ensuring
that no single feature dominates the regression model.

[9]: from [Link] import StandardScaler

# Separate features (X) and target (y)


X = [Link]("Price", axis=1) y =
df["Price"]

# Standardize features scaler =


StandardScaler()
X_scaled = scaler.fit_transform(X)

print("Shape of X:", X_scaled.shape) print("Shape of y:",


[Link])
Shape of X: (5000, 5)

Shape of y: (5000,)

5
[11]: from sklearn.model_selection import train_test_split

# Split data (80% train, 20% test)


X_train, X_test, y_train, y_test = train_test_split( X_scaled, y,
test_size=0.2, random_state=42
)

print("Training set size:", X_train.shape) print("TesBng set size:",


X_test.shape)

Training set size: (4000, 5)


TesBng set size: (1000, 5)
0.0.2 Question 3: Linear Regression Implementation

We selected Price as the continuous dependent variable.


Steps performed:
- Generated a correlation heatmap to observe relationships between features and target.
- Trained a Linear Regression model on the training set.
- Plotted Predicted vs Actual values with a 45° reference line.
- Visualized model errors using a Residual plot.
- Evaluated performance using metrics: R2 Score, MAE, MSE, RMSE.

[12]: import seaborn as sns import


[Link] as plt

# Correlation heatmap plt.figure(figsize=(8,6))


corr = [Link]() # correlation matrix
[Link](corr, annot=True, fmt='.2f', cmap='coolwarm', square=True)
[Link]("CorrelaBon Heatmap (including Price)") [Link]()

6
[13]: from sklearn.linear_model import LinearRegression

# Create and train the model model =


LinearRegression() model.fit(X_train,
y_train)

# Make predictions on test data


y_pred = [Link](X_test)

print("Model trained successfully!") print("First 5


predicBons:", y_pred[:5])

Model trained successfully!

First 5 predicBons: [1308587.92699753 1237037.22949428 1243429.34030687

1228900.21360379

7
1063320.9071082 ]

[14]: import [Link] as plt

plt.figure(figsize=(6,6))
[Link]]er(y_test, y_pred, s=20, alpha=0.6) mn =
min(y_test.min(), y_pred.min()) mx = max(y_test.max(),
y_pred.max())
[Link]([mn, mx], [mn, mx], color='red', lw=2) # 45-degree reference line [Link]("Actual
Price") [Link]("Predicted Price")
[Link]("Predicted vs Actual (Linear Regression)") [Link]()

8
[15]: import seaborn as sns

# Calculate residuals
residuals = y_test - y_pred

# Residual plot plt.figure(figsize=(10,4))

# Scatter plot of residuals vs predicted [Link](1,2,1)


[Link]]er(y_pred, residuals, s=20, alpha=0.6) [Link](0,
color='red', linestyle='--') [Link]("Predicted Price")
[Link]("Residuals (Actual - Predicted)") [Link]("Residuals
vs Predicted")

9
# Histogram of residuals
[Link](1,2,2)
[Link](residuals, kde=True) [Link]("Residuals")
[Link]("Residuals DistribuBon")

plt.Bght_layout() [Link]()

[16]: from sklearn import metrics


import numpy as np

# Calculate evaluation metrics r2 = metrics.r2_score(y_test,


y_pred) mae = metrics.mean_absolute_error(y_test, y_pred)
mse = metrics.mean_squared_error(y_test, y_pred) rmse =
[Link](mse)

print("Model EvaluaBon Metrics:")


print("R² Score:", round(r2, 4))
print("MAE:", round(mae, 2)) print("MSE:",
round(mse, 2)) print("RMSE:", round(rmse,
2))

Model EvaluaBon Metrics:


R² Score: 0.918
MAE: 80879.1
MSE: 10089009300.89
RMSE: 100444.06

10
0.1 Question 4: Polynomial Regression

• Apply Polynomial Regression (degree 2 and 3) on the dataset.

• Plot regression curves to compare with Linear Regression.

• Evaluate performance using R², MAE, MSE, RMSE.

• Conclude whether Polynomial Regression improves generalization or causes overfitting.

[19]: # Degree 2 polynomial poly2 =


PolynomialFeatures(degree=2) X_train_poly2 =
poly2.fit_transform(X_train)
X_test_poly2 = [Link](X_test)

# Degree 3 polynomial poly3 =


PolynomialFeatures(degree=3) X_train_poly3 =
poly3.fit_transform(X_train)
X_test_poly3 = [Link](X_test)

[20]: # Train Polynomial Regression (Degree 2) lin_reg2 =


LinearRegression() lin_reg2.fit(X_train_poly2, y_train)

# Predict on training and test set


y_train_pred2 = lin_reg2.predict(X_train_poly2)
y_test_pred2 = lin_reg2.predict(X_test_poly2)

# Train Polynomial Regression (Degree 3) lin_reg3 =


LinearRegression() lin_reg3.fit(X_train_poly3, y_train)

# Predict on training and test set


y_train_pred3 = lin_reg3.predict(X_train_poly3)
y_test_pred3 = lin_reg3.predict(X_test_poly3)

11
mse = mean_squared_error(y_true, y_pred) rmse =
[Link](mse)

print(f"{model_name} Performance:")
print(f" R² Score: {r2:.4f}") print(f"
MAE: {mae:.2f}") print(f" MSE:
{mse:.2f}") print(f" RMSE:
{rmse:.2f}")
print("-"*40)

# Evaluate Degree 2 Polynomial Regression evaluate_model(y_test, y_test_pred2,


"Polynomial Regression (Degree 2)")

# Evaluate Degree 3 Polynomial Regression


evaluate_model(y_test, y_test_pred3, "Polynomial Regression (Degree 3)")
Polynomial Regression (Degree 2) Performance:
R² Score: 0.9179
MAE: 80886.67
MSE: 10099268148.79
RMSE: 100495.12

Polynomial Regression (Degree 3) Performance:


R² Score: 0.9174
MAE: 80942.67
MSE: 10162847726.45
RMSE: 100810.95

[ ]:

12
13

You might also like