Module 3 Test: Statistics and Math for ML
This is your progress tracker . 20 questions: 5 Easy, 10 Moderate, 5 Hard . Solve all the questions yourself first (work out the output in your head before running the code in Colab), then check with the Answer Key.
| Score | Meaning |
|---|---|
| 18 to 20 | Excellent! Your grip on Module 3 is strong. |
| 12 to 17 | Good. Re-read the chapters where you made mistakes. |
| Below 12 | Do not worry. Re-read Chapters 9 to 11 and run the code yourself, then take the test again. |
EASY
Q1. When data has some very large, unusual values (outliers), which measure is better for describing the centre?
A) Mean B) Median C) Range D) Variance
EASY
Q2. In what range can the Pearson correlation (r) value lie?
A) 0 to 1 B) -1 to +1 C) 0 to 100 D) -100 to +100
EASY
Q3. In a hypothesis test, what do we do if the p-value is smaller than alpha (0.05)?
A) Accept H0 as correct B) Reject H0 C) Reduce the sample size
D) Run the test again for no reason
EASY
Q4. In simple linear regression y = b0 + b1 x, what does the slope (b1) mean?
A) The value of y when x = 0 B) The change in y for a 1-unit increase in x C) The total count of the data D) The sum of errors
EASY
Q5. What is the small part of the population from which data is collected called?
A) Parameter B) Sample C) Variable D) Outlier
MODERATE
Q6. What will be the output of this code (population standard deviation)?
import numpy as np
print(np.std([2, 4, 4, 4, 5, 5, 7, 9]))
A) 2.0 B) 2.14 C) 4.0
D) 5.0
MODERATE
Q7. What will be the output of this code (Pandas default std)?
import pandas as pd
s = pd.Series([2, 4, 4, 4, 5, 5, 7, 9])
print(round(s.std(), 2))
A) 2.0 B) 2.14 C) 4.0 D) 1.9
MODERATE
Q8. What will be the output of this code?
import math
print(math.comb(5, 2))
A) 20 B) 10 C) 5 D) 25
MODERATE
Q9. A student scored 85 marks. The class mean is 70 and the std is 5. What is this student's z-score?
print((85 - 70) / 5)
A) 1.0 B) 2.0 C) 3.0 D) 15.0
MODERATE
Q10. What will be the output of this code?
from scipy.stats import norm
print(round(norm.cdf(0), 2))
A) 0.0 B) 0.5 C) 0.95 D) 1.0
MODERATE
Q11. What is the correct meaning of a 95% confidence interval?
A) The true mean is in this range with 95% probability (for one single interval)
B) If this method is repeated again and again, about 95% of the intervals will capture the true mean
C) 95% of the data is in this range
D) The sample mean is 95% correct
MODERATE
Q12. What is the correct meaning of Type I error?
A) H0 was true but we rejected it (false alarm)
B) H0 was false but we did not reject it
C) The sample size was very small
D) The p-value came out negative
MODERATE
Q13. What will be the output of this code (rounded)?
import numpy as np
b1, b0 = np.polyfit([1, 2, 3, 4], [3, 5, 7, 9], 1)
print(round(b1, 1), round(b0, 1))
A) 1.0 2.0 B) 2.0 1.0 C) 2.0 3.0 D) 3.0 2.0
MODERATE
Q14. 20 students took a test before and after a training. Did the training change their marks? Which test is correct?
A) Independent two-sample t-test B) Paired t-test C) Chi-square test
D) One-way ANOVA (on a single group)
MODERATE
Q15. In simple regression the correlation is r = 0.8. What will R-squared be?
print(round(0.8 ** 2, 2))
A) 0.8 B) 0.64 C) 0.4 D) 1.6
HARD
Q16. A disease affects 2% of people. The test detects the sick correctly 90% of the time and wrongly reports 10% of healthy people as positive. What is the probability of being sick given a positive result (Bayes)? What will be the output of this code?
p_d, sens, fp = 0.02, 0.90, 0.10
p_pos = sens * p_d + fp * (1 - p_d)
print(round(sens * p_d / p_pos, 2))
A) 0.90 B) 0.16 C) 0.50 D) 0.02
HARD
Q17. What is the probability of exactly 2 heads in 3 coin tosses (Binomial)? What will be the output of this code?
from scipy.stats import binom
print(round(binom.pmf(2, 3, 0.5), 3))
A) 0.250 B) 0.375 C) 0.500 D) 0.667
HARD
Q18. What will be the output of this code?
from scipy import stats
r = stats.ttest_1samp([5, 6, 7, 8, 9], 5)
print(round(r.statistic, 2), r.pvalue < 0.05)
A) 0.0 False B) 2.83 True C) 2.83 False
D) 7.0 True
HARD
Q19. What will be the output of this code (Spearman and Pearson, rounded)?
from scipy import stats
import numpy as np
x = np.arange(1, 11)
y = x ** 2
print(round(stats.spearmanr(x, y)[0], 2), round(stats.pearsonr(x, y)[0], 2))
A) 1.0 1.0 B) 1.0 0.97 C) 0.97 1.0 D) 0.97 0.97
HARD
Q20. The sample std is s = 10. If the sample size is increased from n = 25 to n = 100, what happens to the Standard Error (s / sqrt(n))? Output:
import numpy as np
print(10 / np.sqrt(25), 10 / np.sqrt(100))
A) 2.0 and 1.0 (half) B) 2.0 and 0.5 (a quarter) C) 4.0 and 1.0 D) SE does not change
MODULE 3 TEST
Module 3 Answer Key (Answers and explanations)
| # | Level | Answer | Explanation |
|---|---|---|---|
| 1 | Easy | B) Median | The median is the middle value and is less affected by outliers. The mean is pulled toward outliers. |
| 2 | Easy | B) -1 to +1 | Correlation always lies between -1 (a perfect inverse relationship) and +1 (a perfect direct relationship). |
| 3 | Easy | B) Reject H0 | p-value < alpha means the data is very unusual under H0, so H0 is rejected. |
| 4 | Easy | B) The change in y for a 1-unit increase in x | The slope tells how much y changes (on average) for a one-unit increase in x. The intercept is the value of y at x = 0. |
| 5 | Easy | B) Sample | The small chosen part of the population is called a Sample. A number computed from a sample is called a statistic. |
| 6 | Moderate | A) 2.0 | Mean = 5, variance (divided by n) = 32 / 8 = 4, so std = sqrt(4) = 2.0. NumPy's default is ddof=0. |
| 7 | Moderate | B) 2.14 | Pandas' default is ddof=1 (sample std): sqrt(32 / 7) = 2.14. This differs from NumPy's ddof=0 result. |
| 8 | Moderate | B) 10 | Choosing 2 out of 5 as combinations = 5! / (2! x 3!) = 10. (A permutation would be 20, where order matters.) |
| 9 | Moderate | C) 3.0 | z = (x - mean) / std = (85 - 70) / 5 = 3.0. This is 3 std above the mean (very unusual, probably an outlier). |
| 10 | Moderate | B) 0.5 | The standard Normal is symmetric and 0 is its centre, so the portion below 0 is 50% (0.5). |
| 11 | Moderate | B) If this method is repeated again and again, about 95% of the intervals will capture the true mean | The confidence level of a CI is the long-run success percentage of the method, not the probability for one chosen interval. |
| 12 | Moderate | A) H0 was true but we rejected it (false alarm) | Type I = false positive: there was actually no difference but the test reported one. Its probability is alpha. |
| 13 | Moderate | B) 2.0 1.0 | The data lies exactly on y = 1 + 2x: slope 2.0 and intercept 1.0. |
| 14 | Moderate | B) Paired t-test | The same people were measured twice (before/after), so a paired t-test is correct. |
| 15 | Moderate | B) 0.64 | In simple linear regression R-squared = the square of r = 0.64, i.e. the model explains 64% of the spread in y. |
| 16 | Hard | B) 0.16 | p_pos = 0.018 + 0.098 = 0.116. Posterior = 0.018 / 0.116 = 0.155 (0.16). Because the disease is in few people (base rate), the chance stays low even after a positive test. |
| 17 | Hard | B) 0.375 | C(3, 2) x 0.5<sup>2</sup>x 0.5<sup>1</sup>= 3 x 0.125 = 0.375. |
| 18 | Hard | B) 2.83 True | Sample mean = 7, std = 1.58, SE = 0.707, t = (7 - 5) / 0.707 = 2.83. With n = 5 this gives p < 0.05, so True. |
| 19 | Hard | B) 1.0 0.97 | y = x<sup>2</sup>always increases, so Spearman (on ranks) is exactly 1.0. But the relationship is not a straight line, so Pearson is slightly below 1 (0.97). |
| 20 | Hard | A) 2.0 and 1.0 (half) | Multiplying n by 4 halves SE (1 / sqrt(4) = 1/2): 10/5 = 2.0 and 10/10 = 1.0. SE falls with the square root of n, not directly. |
Your score: ______ / 20