Chapter 2 · Section 4 practice

Feature Pipelines and Research Applications

0 of 8 answered

These questions review the examples in the companion slides. Choose an answer, then check the explanation.

Question 1

Why can ordinary steps and a short chain give the same table?

  • They perform the same operations on the same input in the same order
  • A chain always removes missing values
  • Chaining changes the formula
  • A chain ignores the row index

Question 2

Which write updates the selected cells of the original table?

  • df.loc[mask]["flag"] = True
  • df[mask]["flag"] = True
  • df.copy()["flag"] = True
  • df.loc[mask, "flag"] = True

Question 3

Why use groupby("ticker")["ret"].shift(1)?

  • To average all dates
  • To remove every missing row
  • To keep each stock’s previous return within that stock
  • To copy the last stock’s return into the next stock

Question 4

The first row is January. A three-month window on returns shifted by one first becomes complete in which month?

  • March
  • April
  • January
  • February

Question 5

What should happen to the missing initial historical-window values?

  • Retain them as missing and report feature coverage
  • Replace them by zero return
  • Borrow the next stock’s history
  • Duplicate the first valid value backwards

Question 6

Two same-date sector momentum values are 0.10 and 0.30. What is the first demeaned value?

  • 0.10
  • 0.20
  • 0.40
  • -0.10

Question 7

A sector has only one observed momentum value. Why is its sample z-score missing?

  • Its price must be missing
  • All singleton groups are data errors
  • The sample standard deviation is undefined with one observation
  • The stock must have zero risk

Question 8

What does a momentum percentile rank describe?

  • Its portfolio weight
  • A stock’s relative ordering within the same-date comparison
  • Its probability of a positive future return
  • Its predicted return in percent