Exam 3 - Chapter 8 - Predictive Analytics

Question 1

Define predictive analytics:

Answer 1

Analytics performed to provide foresight by identifying patterns in historical data to judge likelihood of future events.

Question 2

What is the main difference between:

• descriptive and diagnostic analytics

and

• predictive and prescriptive analytics

Answer 2

Descriptive and diagnostic

Generally report known facts

Predictive and prescriptive

Provide more probabilistic model

Question 3

What are the three groups of predictive analytics?

Answer 3

• Classification
• Regression
• Time series analysis

Question 4

Define classification of predictive analytics?

What differentiates classification from regression?

Answer 4

Analytics technique used to separate or classify a sample into two or more groups.

• The output (y-variable) of classification is categorical
• The output variable for regression is numerical

Question 5

What are a few accounting based questions that classification can answer?

Answer 5

• Bankruptcy classification
• Loan extension classification
• Fraud/no fraud
• Going concern/no going concern

Question 6

What is a technique used in bankruptcy classification of predictive ananlytics?

Answer 6

Altman’s Z:

Categorizes businesses into classes that determine the risk of bankruptcy based on five business ratios

Question 7

What are the five business ratios used in Altman’s Z test?

What do these ratios measure?

Answer 7

Working Capital/Total assets

Retained earnings/total assets

Earnings before interest and taxes/total assets

FV of SH equity/ BV total debt owed

Sales / Total assets

Question 8

The equation for altman’s bankruptcy analysis is:

Answer 8

Z = 1.2x₁ + 1.4 x₂ + 3.3 x₃ + 0.6x4 + 1.0x₅

If Z < 1.8 : significant risk of bankruptcy

If Z > 3.0 : no risk of bankruptcy

Question 9

How does loan extension fall under classification of predictive analytics?

Answer 9

Uses independent variables to classify loan acceptance or loan rejection of customers

• Credit history
• Income
• Loan amount
• Employment length
• Debt-to-income ration
  *

Question 10

What factors increase the risk of fraud classification when classifying a company for fraud/no fraud?

Answer 10

1. Increase in receivables from prior period
2. Decline in gross margin
3. Decline in asset quality index (more LT assets)
4. Increase in sales growth
5. Decrease in depreciation expense
6. Decrease in SG&A expenses
7. Increase in debt
8. Higher total accruals to total assets (more profit on account)

Question 11

What is regression analysis for predictive analytics?

Answer 11

Method of determining relationship between two sets of variables when one variable is dependent on another independent variable.

Question 12

How can regression analysis be used for predictive analytics?

Answer 12

• How are firms costs dependent on level of production?
• What are the cost drivers for overhead costs?
• What dependent variables do lendors use to set interest rates?
• What is relationship between investment risk and returns

Question 13

What are base rates and how do they impact predictions?

Answer 13

Base Rates

Probability of an event occuring based on historical average

• Base rates can be used to gauge reasonableness of predictions

Question 14

What is a base rate fallacy?

Answer 14

Base rate fallacy

Prediction places too little weight on base rates and uses different information to base prediction

Question 15

What is time series analysis?

What is persistence in time series analysis?

Answer 15

Time series analysis

Technique used to predit future values based on trends of past values of the same variable

Persistence: Continuity and stability of financial statement variables (will trends continue)?

Question 16

Persistence can be calculated using this regression analysis:

Performance_t+1 = β₀ + β₁ (Performance)

What variable determines persistence?

Answer 16

β₁

The closer β1 is to 1.00, the more persistent the variable is

Question 17

Dechow and Schrand (2004) looked at the persistence of various financial statement variables over the 1987–2002 time period.

What do these figures indicate?

Answer 17

• Generally high levels of persistence indicate that prior year variables are reasonable predictors
• Operating income’s persistence > Cash flows from operating activities: accrual accounting provides more consistent information than cash

Question 18

Why is hypothesis testing important in predictive analytics?

Answer 18

Can create hypotheses and perform statistical tests (t-tests) to determine if scenarios are statistically significant

Question 19

What is machine learning?

Answer 19

Machine learning

Ability of a computer to automatically learn on its own (without programming)

Question 20

How is machine learning applicable to accounting and predictive analytics?

Answer 20

Robotic process automation (RPA)

Software programs that automate certain repeatable, accounting tasks

Can be used to predict which tasks require additional intervention by accountants