Scribd
Upload
15 views

answer_key

Uploaded by

asher2020zhang
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
15 views

answer_key

Uploaded by

asher2020zhang
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
You are on page 1/ 1

INSTRUCTOR ANSWER KEY

-----------------

Test Case 1: 3000 sq ft, 4 bedrooms


Actual Expected Price Range: $580,000 - $620,000
Best Performing Model: Random Forest
Reasoning:
1. The relationship between house features and price is likely non-linear
2. Random Forest handles the combination of discrete (bedrooms) and continuous
(square feet) features well
3. The dataset shows a clear pattern of price increasing with size and bedrooms
4. Predictions will likely cluster around $600,000

Test Case 2: 1600 sq ft, 3 bedrooms


Actual Expected Price Range: $330,000 - $360,000
Best Performing Model: Linear Regression
Reasoning:
1. This house is close to the center of the dataset
2. Linear relationship is more apparent for mid-range houses
3. The features are very close to the mean of the training data
4. Predictions will likely be most consistent with linear regression
5. Expected price: Around $340,000

Test Case 3: 2700 sq ft, 4 bedrooms


Actual Expected Price Range: $530,000 - $570,000
Best Performing Model: Random Forest
Reasoning:
1. Larger house with 4 bedrooms
2. Requires capturing non-linear price scaling
3. At the upper end of the dataset, linear models may start to deviate
4. Random Forest can better capture the nuanced price increase
5. Expected price: Around $550,000

Test Case 4: 1200 sq ft, 2 bedrooms


Actual Expected Price Range: $240,000 - $260,000
Best Performing Model: Linear Regression
Reasoning:
1. Smaller house with fewer bedrooms
2. Located at the lower end of the dataset
3. Linear relationship is more predictable for smaller houses
4. Simple features make linear regression effective
5. Expected price: Around $250,000

Overall Observations:
- Random Forest performs best for larger, more complex houses
- Linear Regression works well for houses near the dataset's center
- Neural Network may show high variance due to small dataset
- SVR can be unpredictable with limited training data

Key Learning Points:


1. No single model is perfect for all scenarios
2. Model selection depends on data characteristics
3. Always validate predictions against domain knowledge
4. More data leads to more reliable predictions

You might also like