Exam 1 Review
What is what-if analysis
An analysis of what would happen to the solution derived from a decision model if the values assigned to one or more of the parameters were to be changed to other plausible values. Also called post-optimality analysis (or sensitivity analysis)
What is descriptive analytics
Analyzing data to create descriptions of what as happened so far
Managers make decisions based off of what
Answer
What are techniques used in prescriptive analytics
Answer
What is the linear regression trying to do between the dependent and independent variables
Approximating the relationship between the dependent variable and the independent variable(s) by a straight line
What else does the study team do once it has a problem defined
Are there some metrics for measuring the performance to date thatmanagement would like illuminated through descriptive analytics and datavisualization?• Are there any predictions of the behavior of prospective customers that wouldbe of interest to management? If so, predictive analytics can be used todevelop these predictions.• Does management want certain prescriptions of what should be done in thefuture by using prescriptive analytics
What are the two different predictive analytics models
Categorical (classification models) and Prediction models (numerical outcomes)
What is a collection of data
Dataset
What is a model
Decision model - mathematical model used for decision making that prescribes what should be done in the future Prediction Model - Model that predicts a numeric outcome Classification model - Attempts to predict yes or no outcome are representations of real objects or situations andcan be presented in various forms
What is a data dictionary
Describes the meaning of each variable for the record
What is data exploration
Discovery through numerical summaries and visualizations A widely used process that does this is the Knowledge Discovery inDatabases (KDD) process This process commonly leads to calculating performance metrics(numerical measures of performance) that provide an accurate pictureof what has been happening recently and how this compares with otherperiods Summary statistics are useful for quantifying various characteristics ofindividual variables and the relationship between pairs of variables.• Another method of exploring the data and the relationships betweenvariables is to visualize the data by using helpful graphics.18
In prescriptive analytics what is the step after cleaning the data
Explore the data
After cleaning the data what is the next step
Explore the data (descriptive)
What is the process of ETL
Extract, transform, load
A mathematic equation allows a manager to make a decision for both quantitative and qualitative T or F
F
All variables in a dataset have to be numbers T or F
F
Descriptive, prescriptive, and predictive have the same goal (T or F)
F
Managers make decisions make decisions based solely on quantitative only (T or F)
F
T or F All prediction models uses data to produce a Y or N outcome
F
T or F Prescriptive analytics involves more extensive problem definition than predictive analytics
F
T or F The KNN algorithm predicts a customer's actions based on upon the actions of those who live in the house next door
F
T or F Predictive analytics is a three step process
F, it is 5 steps STEP 1: Develop a model. STEP 2: Partition the data.S TEP 3: Test and refine the current model. STEP 4: Repeat steps 1−3 for several models, and then choose the best. STEP 5: Implementation
What is the first step in prescriptive analytics
Formulate a decision model
What is the first step in descriptive analytics
Gather and organize relevant data
What is the final step in predictive analytics
Implementation
What is the Delphi method
A judgmental forecasting method that uses input from a panel of experts in different locations Experts fill out a questionnaire and in the next round they are provided answers to narrow down the conclusions from the experts
What is the averaging method in forecasting
A method that uses the average of all the past observations from a time series as a forecast of the next value.
What is the Naive Bayes classifier?
A numerical value can be converted into a categorical variable by binning where the numerical values are divided into bins.
What is the smoothing constant in exponential smoothing
A parameter of the exponential smoothing forecasting method that gives the weight to be placed on the last value in the time series controls the weight given to the most recent observation compared to past observations when generating forecasts. It determines how much the forecast is influenced by new data versus historical data. Closer to 1, the MORE heavy it will be on the recent data
What is training partitioning
A portion of the available historical data for a predictive analytics model that is used to develop ("train") the model
What is validation partitioning
A portion of the available historical data for a predictive analytics model that is used to validate the model after a training partition is used to develop the model.
What is momentum
The movement over the last X days, moving average
What is a predictor record
When making a prediction about the future behavior of a new customer (or another entity), that new customer and its characteristics (the predictor variables) are referred to as the predictor record
What are the two different types of partitioning
Training and validation partitioning
What is a gradual, long term movement called
Trend (The average change from one time series value to the next if the current pattern continues.)
T or F Data exploration often makes use such as scatter plots and pie charts
True
T or F Multiple linear regression can use more than 1 predictor variable
True
What does a KNN algorithm do
Use the outcomes (whether the households had solar installed or not) of these khistorical records to predict the likely outcome for the predictor record
What is done in exponential smoothing
places the greatestweight on the last value in the time series and then progressivelysmaller weights on the older values More sophisticated version of the MA method
What do square nodes represent in a classification tree
represent a region that has not beensplit further
What do the circular node represent on a classification tree
represents a split based upon a dividingline for one predictor variable
What is the salesforce composite method
A judgmental forecasting method that aggregates the sales forecasts of the salesforce from their various regions Bottom up approach
What is a regression tree
A close variant of a classification tree that is used when dealing with numerical outcomes
A regression tree does what
A close variant of a classification tree that is used when dealing with numerical outcomes.
What is the seasonal factor
A factor for any period of a year that measures how that period compares to the overall average for an entire year
What is the most common type of decision model
Linear programming
What is the primary method for causal forecasting
Linear regression
What is naive forecasting and what type of value does it have to be
Looking at the last value and using that as the forecast Has to be numeric
What are homogenous regions
Making the groups as much the same as possible Each node on a classification tree should become more homogenous
What does multiple linear regression utilize
Multiple independent variables
Are forecasts always perfect
No
Are optimizing and satisfying the same thing
No
Is KNN limited to 2 predictor variables
No
What are types of decision models
Non-linear programming model linear programming model Prediction and classification models
What are common methods to rescale data
Normalize the data - Rescaling each type of data so that its values range from 0 to 1 Standardize the data - Rescaling each type of data so that it measures the number of standard deviations above or below the mean for a standard normal distribution
Management science is based off of what field
Operations research
What is the goal of constructing a classification tree
Predict categorical outcome based on predictor values
What is a predictive model that is used to estimate future values
Prediction model
What are the two kinds of prediction analytics
Predictive and categorical
What do managers base their decisions on
Quantitative and qualitative factors
What does R square do
R-squared measures how well the independent variables explain the variation in the dependent variable
What are the steps in the prescriptive analytics process
STEP 1: Formulate a decision model STEP 2: Learn how to derive solutions from the model STEP 3: Test the model STEP 4: Prepare to apply the model STEP 5: Implementation
What are the steps in the descriptive analytics process
STEP 1: Gather and organize relevant data STEP 2: Clean the data STEP 3: Explore the data STEP 4: Communicate performance information using data visualization
What is a data warehouse
System that aggregates data from many sources into a single collection
Business analytics draws upon a variety of quantitative decision sciences for techniques that are useful for analyzing data (T or F)
T
Management science is a discipline that attempts to aid managerial decisions by applying the scientific approach to quantitative T or F
T
Prescriptive analytics uses analytical techniques to determine the best option for future decisions T or F
T
T or F Before gathering data, a project team should have a clear definition of the study objectives and the evaluation criteria
T
T or F Categorical variables can only take on a small number of possible values
T
What is the mean squared error (MSE)
The average of the square of the forecasting errors
What are some details about the study team
The study team normally works in an advisory capacity.They advise management (often one key decision maker)
What does Big Data refer to
Where enormous and increasing amounts of transactional data commonly are available for analysis
Can you use a prediction model to get a numeric outcome
Yes
Does prescriptive analytics involve a decision model and a course of action
Yes
Should you rescale the data if you are using KNN
Yes
What kind of discipline is management science
a discipline that attempts to aidmanagerial decision making by applying a scientificapproach to managerial problems that involve quantitativefactors
What is multiple linear regression
a model with more than one independent variable Model for predicting numerical outcome to determine the relationship between the numerical outcome and the MULTIPLE predictor variables
Descriptive analytics is a process of using data to do what
aids in describing, demonstrating, or helpfully summarizing data points so those patterns may develop that satisfy all of the conditions of the data
Business Analytics is a field which....
aids managerial decision making through the use of data.
What does business analytics draw upon
all the other quantitative decision sciences, such as mathematics,statistics, computer science, data science, industrial engineering, etc
What do prescriptive analytics do
analyze data and content to recommend the optimal course of action or strategy moving forward
What is the last value forecast method
gnores all data pointsin a time series except the last one.Forecast = Last value
What is Mean absolute deviation (MAD)
the average of the absolute values of deviations of observations in a data set from its mean The average forecasting error
What does the study team do
the first order of business is to study the relevant system and then toconsult with management to develop a well-defined statement of the problem
What is business analytics
the scientific process of transforming data into insight for making better decisions
What does predictive analytics do
uses data to find hidden relationships that will help us predict what's going to happen in the future
When is a classification tree most useful and appropriate
using a scatter plot is used to illustrate the process of forming regions to build intuition about what is happening but a classification tree is most commonly used for this A classification tree is more versatile and amenable compared to a scatterplot