Predictions, decisions and recommendations from models
Name:
Date:
Class:
A
Quick Questions
Shade in the A or B circle for each question.
1. A model uses past rainfall data to say it will probably rain tomorrow. Which use of a model is this?
2. A streaming app suggests three shows you might like based on what you watched before. Which use of a model is this?
3. Which statement best describes what a model does with data?
4. A school canteen uses a model of past sales to decide how many pies to order for Friday. Which use of a model is this?
5. An online shop shows 'Customers who bought this also bought...' under a backpack. What is the model doing?
B
Match the Pairs
Draw a line connecting each item on the left to its match on the right.
| Left column | Right column | |
|---|---|---|
| 1=FORECAST(A9,B2:B8,A2:A8) | AFORECAST is missing the known_xs argument. It needs the new x, the known y values and the known x values. | |
| 2=IF(C2>=50,"Reorder","OK") | B=FORECAST(8,B2:B7,A2:A7) | |
| 3A model predicts what will happen next from past data | CUses a rule from a model to support a stock decision | |
| 4A model recommends options to users, such as suggested products | DPredictions from a model are estimates based on past data, not certain facts, so they can be wrong. | |
| 5A model always gives a certain, guaranteed prediction of the future. | ESELECT name FROM products WHERE category='Tech' ORDER BY price; | |
| 6=FORECAST(A9,B2:B8) | FUses a model built from past data to predict a future value |
C
Definitions
model
D
Fill in the Blanks
Complete each line of code by filling in the missing parts.
1.
=________(6,B2:B6,A2:A6)
2.
=__(B2>100,"Reorder","Wait")
3.
=IF(B2__80,"Try extension","Revise")
4.
______ name FROM products WHERE stock > 0;
5.
=IF(C2_10,"Order more","Enough stock")
E
Spot the Error
Shade in the letter of the line that contains an error.
Question 1
A
sales ← LOAD(history)
B
model ← TRAIN(next)
C
next ← PREDICT(model, 7)
D
DISPLAY next
Question 2
A
=FORECAST(A8,A2:A7,B2:B7)
B
=IF(B8>100,"Order","Wait")
C
=B8-B7
Question 3
A
SELECT name, price
B
FROM products
C
WHERE stock = 0
D
ORDER BY price ASC
E
LIMIT 3
Question 4
A
demand ← 120
B
stock ← 100
C
IF demand < stock THEN
D
DISPLAY "Order more"
E
END IF
Question 5
A
best ← 0
B
FOR each item IN products
C
IF item.score > pick THEN
D
best ← item.score
E
pick ← item
F
DISPLAY pick
Answer Key
Predictions, decisions and recommendations from models
Section A: Quick Questions
Q1
A — Prediction
Q2
B — Recommendation
Q3
A — It uses data to predict outcomes, support decisions and recommend options
Q4
A — Supporting a decision
Q5
B — Recommending suggested products
Section B: Match the Pairs
Q1
=FORECAST(A9,B2:B8,A2:A8) → Uses a model built from past data to predict a future value
Q2
=IF(C2>=50,"Reorder","OK") → Uses a rule from a model to support a stock decision
Q3
A model predicts what will happen next from past data → =FORECAST(8,B2:B7,A2:A7)
Q4
A model recommends options to users, such as suggested products → SELECT name FROM products WHERE category='Tech' ORDER BY price;
Q5
A model always gives a certain, guaranteed prediction of the future. → Predictions from a model are estimates based on past data, not certain facts, so they can be wrong.
Q6
=FORECAST(A9,B2:B8) → FORECAST is missing the known_xs argument. It needs the new x, the known y values and the known x values.
Section C: Definitions
Q1
model — A model uses data to predict what will happen, support decisions and recommend options to users, such as suggested products.Section D: Fill in the Blanks
Q1
=FORECAST(6,B2:B6,A2:A6)Q2
=IF(B2>100,"Reorder","Wait")Q3
=IF(B2>=80,"Try extension","Revise")Q4
SELECT name FROM products WHERE stock > 0;Q5
=IF(C2<10,"Order more","Enough stock")Section E: Spot the Error
Q1
B — A model is trained on past data, so it must be TRAIN(sales). The variable next does not exist yet. It is the prediction that comes after training.
Q2
A — FORECAST takes x, then the known y values, then the known x values. The sales column (B) must come before the month column (A): =FORECAST(A8,B2:B7,A2:A7).
Q3
C — A recommendation should only suggest items that are available. The condition should be stock > 0. With stock = 0 the query would suggest items that are sold out.
Q4
C — The decision to order more should be made when predicted demand is greater than stock. It should be demand > stock, not demand < stock.
Q5
C — Each item's score is compared with the best score so far, so it should be item.score > best. The variable pick holds an item, not a score.