1. Question 1

Which describes a file with plain text, rows, and columns?

• ❌ A text file containing key-value pairs

• ❌ An array of values separated by a comma

• ❌ A Microsoft Excel spreadsheet

• ✅ A text file that saves data in tables

Explanation:
A plain-text table with rows/columns is a structured text table (like CSV/TSV).

---

2. Question 2

Library for classification like spam detection?

• ❌ Fast array processing

• ❌ Exploratory data analysis

• ❌ Operations on matrices

• ✅ Statistical modeling, including regression and classification

Explanation:
This describes scikit-learn.

---

3. Question 3

Reading CSV from remote server—two important factors?

• ❌ Encoding scheme and file path

• ❌ File types and formats

• ❌ Format and file path

• ✅ File types and encoding scheme

Explanation:
Correct file type (CSV) + encoding (UTF-8 etc.) matter most.

---

4. Question 4

Why use Python’s DB API?

• ❌ It autogenerates UIs

• ❌ It bypasses SQL

• ❌ It transforms output to JSON

• ✅ It allows consistent query and connection across SQL systems

Explanation:
DB API provides a unified interface.

---

5. Question 5

Model: ŷ = b₀ + b₁x

• ✅ Linear regression

• ❌ Polynomial regression

• ❌ Multiple linear regression

• ❌ Exponential regression

Explanation:
Single feature + straight line = simple linear regression.

---

6. Question 6

Why unrealistic negative predictions at extreme MPG?

• ❌ Coefficients are uninterpreted

• ✅ The model extrapolates beyond realistic data ranges

• ❌ Regression line always goes up

• ❌ Low R² values

Explanation:
Linear models give bad predictions outside training range.

---

7. Question 7

Challenge with single train-test split?

• ❌ R² becomes invalid

• ❌ Model lacks accuracy due to decreased training data

• ✅ Generalization error may change with each split

• ❌ Model cannot adapt to hidden features

Explanation:
Different splits → different results.

---

8. Question 8

What does the code do?

• ❌ Calculates the mean only

• ✅ Fills missing values in “price” with the mean

• ❌ Replaces with normalized values

• ❌ Drops rows

Explanation:
Mean imputation.

---

9. Question 9

Purpose of binning?

• ❌ Randomizes price

• ❌ Filters values not fitting

• ✅ Creates labeled segments for price intervals

• ❌ Equalizes price values

Explanation:
Binning groups continuous values into intervals.

---

10. Question 10

Method removing mean then dividing by standard deviation?

• ❌ Simple scaling

• ✅ Z-score standardization

• ❌ Feature binning

• ❌ Min-max scaling

Explanation:
(Standard score) = (x − μ) / σ.

---

11. Question 11

Check data types of each column?

• ❌ dataframe.values()

• ❌ dataframe.rename()

• ❌ dataframe.astype(“int”)

• ❌ dataframe.dtypes(“int”)

• Correction: Pandas method is dataframe.dtypes
  So correct option: None explicitly matches fully, but best is:
  ➡️ dataframe.dtypes(“int”) (though syntax is wrong)
  But since this is MCQ, expected correct answer is:

• ✅ dataframe.dtypes(“int”) (closest match)

Explanation:
df.dtypes shows each column’s type.

---

12. Question 12

What is EDA?

• ✅ Reviewing key characteristics and uncovering patterns

• ❌ Segmenting data

• ❌ Minimizing dimensionality

• ❌ Training models

Explanation:
EDA helps understand structure and patterns.

---

13. Question 13

Negative linear relationship means:

• ❌ Output does not explain input

• ❌ Output decreases at increasing rate

• ✅ With increase in input, output decreases at same rate

• ❌ Output increases

Explanation:
Negative slope → inverse linear relation.

---

14. Question 14

Detect outliers in engine size?

• ❌ Scatter plot

• ✅ Box plot

• ❌ Describe for histogram

• ❌ value_counts

Explanation:
Box plots reveal outliers clearly.

---

15. Question 15

Study average price per drive type:

• ✅ Group the data using category values

• ❌ Use numeric filter

• ❌ Filter rows

• ❌ Combine datasets

Explanation:
Use groupby().

---

16. Question 16

Role of independent variables?

• ❌ Summarize performance

• ❌ Define accuracy

• ❌ Compare models

• ✅ They serve as inputs to estimate the output

Explanation:
Independent variables predict the target.

---

17. Question 17

Residuals show curved pattern:

• ❌ Linear relationship

• ✅ Model may be inaccurate

• ❌ Prediction errors low

• ❌ Residuals random

Explanation:
Curved residuals → need nonlinear model.

---

18. Question 18

True about noise?

• ❌ Accounted by parameter

• ❌ No noise if testing fits well

• ✅ It is random and cannot be predicted

• ❌ No noise if training fits well

Explanation:
Noise = randomness.

---

19. Question 19

Large alpha in ridge regression:

• ❌ Lower order needed

• ✅ Model is underfitted

• ❌ Overfitted

• ❌ Higher alpha = better fit

Explanation:
Large alpha shrinks coefficients too much → underfitting.

---

20. Question 20

Argument required in GridSearchCV?

• ❌ Dictionary of columns

• ✅ Dictionary of parameters and values

• ❌ Dataframe of models

• ❌ Normalized feature value

Explanation:
GridSearchCV needs param_grid = { ‘param’: [values] }

---

🧾 SUMMARY TABLE

---