BUAL-5650 Exam 2 (Chapters 5-8)

ANN are capable of solving complex problems and provide detailed explanations of their capabilities.

False

An advantage of simulation is that it allows model builders to solve problems with minimal interaction with users or managers.

False

An expected value is computed by dividing the results (i.e., outcomes) by their respective probabilities and subtracting them.

False

Bootstrapping is like the k-fold cross-validation where the k takes the value of 1.

False

CNNs are deep learning architectures that are designed to process only the image data sets.

False

Compute unified device architecture, which enables software developers to use GPUs, was originally designed by Google.

False

Data mining requires specialized data analysts to ask ad hoc questions and obtain descriptive answers quickly from the varied data system.

False

Deep MLP and convolutional networks are specialized for processing a sequential grid of values.

False

Deep learning is a separate branch of computational sciences from AI.

False

Genetic algorithms are a part of global search techniques used to find optimal solutions to complex problems.

False

In NLP, part-of-speech tagging refers to the word sense disambiguation (i.e., words having more than one meaning).

False

In artificial neurons weight terms are adjustable but biased terms are fixed.

False

K in K-means clustering refers to the dimensionality of the input data (i.e., the number of input variables) used in the analysis.

False

One of the disadvantages of simulation modeling is that it usually oversimplifies the underlying real-world system or problem.

False

Representation learning is a subcategory of deep learning that focuses on better representation of complex problems that require unsupervised learning.

False

Search engine optimization is the intentional activity of improving the speed/agility and accuracy of a given search engine.

False

Simulation modeling process starts with the construction of the detail model of the system.

False

Statistics and data mining both look for data sets that are as large and as varied as possible.

False

The main idea in market basket analysis is to predict future customer demand patterns so that an appropriate volume of merchandise would be produced or purchased.

False

The use of simulation models is desirable because they can usually be solved in one pass, without incurring the time and cost of iterations.

False

The value provided by area under the curve (AUC) ranges between -1 and 1 and is the true representation of the prediction accuracy.

False

Theano is a Java library developed by the Deep Learning Group at the University of Montreal.

False

A typical internet search engine is composed of only two main processes, development cycle and response cycle.

True

Bayesian classifiers use probability theory to build classification models based on past occurrences that are capable of placing a new instance into a most probable class.

True

Clickstream analysis does not need users to enter their perceptions of the Web site or other feedback directly, to be useful in determining their preferences.

True

Current use of sentiment analysis in voice of the customer applications allows companies to change their products or services in real time in response to perceived customer opinions.

True

If linear programming can be successfully applied to a problem, the output is usually optimal.

True

In Caffe, everything is done using text files instead of code.

True

In data mining, classification models help in prediction of cases that belong to two or more categories.

True

In genetic algorithms, elitism is used to preserve a few of the best solutions to evolve through the generations.

True

In linear programming formulation, decision variables describe alternative courses of actions.

True

In sentiment analysis, it is hard to classify some subjects such as news as good or bad, but easier to classify others, e.g., movie reviews, in the same way.

True

Making optimal decisions using analytical models is what is called prescriptive analytics.

True

Neural network training is usually done by defining a performance function.

True

Neurons are processing units that perform a set of predefined mathematical operations on the numerical values coming from the input variables or from the other neuron outputs to create and push out its own outputs.

True

Sensitivity analysis seeks to assess the impact of changes in the input data and parameters on the proposed solution.

True

Structured, human-mediated machine-learning approaches have been working fine for rather abstract and formal tasks.

True

Taking a decision under risk is different from taking the decision under uncertainty.

True

Text mining is essentially the same as data mining, except for the addition of information retrieval and extensive preprocessing.

True

The area under the ROC curve is a graphical assessment technique where the true positive rate is plotted on the y-axis and the false negative rate is plotted on the x-axis.

True

The current trend in prescriptive analytics is towards developing and using Web tools and software to access and run modeling software.

True

The growth in the use of deep neural networks has been spurred by advancements in GPU hardware technology.

True

Which broad area of data mining methods partitions a collection of objects into natural groupings with similar features? a) Clustering b) Associations c) Classification d) Visualization

a) Clustering

Which of the following is the most evolved search engine method? a) Cognitive search b) Keyword search c) Contextual search d) Semantic search

a) Cognitive search

Which of the following methods calculates the values of the inputs necessary to achieve a desired level of output? a) Goal seeking b) Utility theory c) What-if analysis d) Multiple regression

a) Goal seeking

Which of the following is a data mining reality? a) If the data accurately reflect the business or its customers, any company can use data mining b) Data mining is not yet viable for mainstream business applications c) Data mining provides instant, crystal-ball-like predictions d) Only those with advanced degrees can do data mining.

a) If the data accurately reflect the business or its customers, any company can use data mining

In handling uncertainty in decision modeling, what does the pessimistic approach do? a) It assumes the worst possible outcome of each alternative will occur and then selects the best of them b) It assumes the worst possible outcome of some alternatives will occur and then selects the best of them c) It assumes the worst possible outcome of one alternative will occur and then avoids it d) It assumes the worst possible outcome of each alternative will occur and then selects the worst of them

a) It assumes the worst possible outcome of each alternative will occur and then selects the best of them

Which of the following is not a method to handle multiple goals? a) Multiple regression model b) Goal programming c) Utility theory d) Expression of goals as constraints, using LP

a) Multiple regression model

Which of the following is not among the factors allowing banks to use more advanced AI based fraud detection systems? a) Security and privacy of data is becoming less of an issue b) Hardware and processors becoming more powerful c) Evolving new learning systems enabled by AI d) Specialized algorithms becoming available

a) Security and privacy of data is becoming less of an issue

Testing the robustness of decisions under changing conditions is an example of ________ analysis. a) Sensitivity b) Heuristic c) Optimality d) Dynamic

a) Sensitivity

Whereas ________ starts with a well-defined proposition and hypothesis, data mining starts with a loosely defined discovery statement. a) Statistics b) Association c) Forecasting d) Clustering

a) Statistics

Which deep learning framework supports TPUs (a type of processor developed by Google)? a) TensorFlow b) Theano c) Caffe d) Torch

a) TensorFlow

In decision-making, fixed factors that affect the result variables but are not manipulated by decision maker are called ________ variables. a) Uncontrollable b) Intermediate c) Result d) Decision

a) Uncontrollable

Which of the following allows extracting useful information from the links embedded in Web documents? a) Web structure mining b) Social media mining c) Text mining d) Web content mining

a) Web structure mining

Which of the following is not an uncontrollable variable? a) Tax rates b) Advertising budget c) Interest rate d) Inflation

b) Advertising budget

Typical use cases for cognitive computing include all of the following, except? a) Fraud detection b) Churn analysis c) Sentiment analysis d) Speech recognition

b) Churn analysis

Which of the following is not among the genetic algorithms' constructs? a) Elitism b) Discretization c) Mutation d) Reproduction

b) Discretization

Environmental scanning is important for all of the following reasons, except? a) Organizational culture is important and affects the model use b) Environments have greater impact on a model than the organization does c) Environmental factors may have created the current problem d) It is critical to identify key corporate decision makers

b) Environments have greater impact on a model than the organization does

The question "What advertising budget is needed to increase market share by 7%?" is a type of ________. a) What-if analysis b) Goal-seeking analysis c) Sensitivity analysis d) Utility modeling

b) Goal-seeking analysis

Sensitivity analysis is important in management support systems for all of the following reasons, except? a) It provides a better understanding of the model and the decision-making situation b) It improves the mathematical optimality of the generated solutions c) It allows flexibility and adaptation to changing conditions d) It permits the manager to input data to increase his/her confidence in the model

b) It improves the mathematical optimality of the generated solutions

Generally, a sequential order of layers must be held between the input and the output layers in this type of network architecture. a) SVM b) MLP c) Knowledge based systems d) RNN

b) MLP

In mathematical programming, of available, the ________ solution is the best, i.e., the degree of goal attainment associated with it is the highest. a) Heuristic b) Optimal c) Dynamic d) Stochastic

b) Optimal

________, also called homonyms, are syntactically identical words with different meanings. a) Antonyms b) Polysemes c) Morphology d) Tokens

b) Polysemes

Which type of neural network is characterized by allowing feedback connections? a) Deep feedforward network b) Recurrent neural network c) Feedforward network d) None of these

b) Recurrent neural network

Which of the following is not among the key reasons data mining is gaining attention in the business world? a) The exponential increase in data processing and storage technologies b) Significant increase in mergers and acquisition in the marketplace c) More intense competition at the global scale d) Recognition of the untapped value hidden in large data sources

b) Significant increase in mergers and acquisition in the marketplace

Clustering partitions a collection of things into segments whose members share which of the following? a) Similar variables b) Similar characteristics c) Similar collection methods d) Similar volume of data

b) Similar characteristics

In neural networks, which of the following is tasked to compute the weighted sums of all input elements entering each processing element? a) Transfer function b) Summation function c) Learning rate d) Backpropagation

b) Summation function

Which deep learning framework is based on LuaJIT library, which is a compiled version of the popular Lua programming language? a) TensorFlow b) Torch c) Caffe d) Theano

b) Torch

LP allocation problems usually display the following characteristics, except? a) The resources are used in the production of products or services b) Unlimited quantity of economic resources available for allocation c) Each activity (product or service) in which the resources are used yields a return d) There are two or more ways in which the resources can be used

b) Unlimited quantity of economic resources available for allocation

Which of the following is not among the factors used to assess a classification model? a) Scalability b) Volume c) Accuracy d) Speed

b) Volume

In data mining, finding products that frequently put into a shopping cart is known as ________. a) Decision trees b) Cluster analysis c) Association rule mining d) Artificial neural networks

c) Association rule mining

Which of the following data mining processes/methodologies is the most comprehensive? a) Six Sigma b) KDD Process c) CRISP-DM d) SEMMA

c) CRISP-DM

The basic idea behind a(n) ________ is that it recursively divides a training set until each division consists primarily of examples from one class. a) Association rule mining b) Stratified random sampling c) Decision tree modeling d) K-fold cross validation

c) Decision tree modeling

In the opening vignette, which of the following was not part of the obtained results? a) Increase in true positives by 50% b) Reduction in false positives by 60% c) Elimination of fraud completely at 15% of branches d) Resources focused on actual cases of fraud

c) Elimination of fraud completely at 15% of branches

GPU hardware is often used for deep neural networks. What does GPU stand for? a) Gated processing unit b) Game programming unit c) Graphics processing unit d) Gated power unit

c) Graphics processing unit

LSTM networks are typically not used for what type of application and/or data types? a) Handwriting recognition b) Machine translation c) Image recognition d) Speech recognition

c) Image recognition

Which of the following is not a component of a linear programming problem? a) Decision variables b) Objective function c) Internal metrics d) Constraints

c) Internal metrics

Why is pretraining a deep MLP network appropriate? a) It increases the duration before global optimum is reached b) It helps in restructuring the input variable space c) It increases the chances of a global optimum being reached d) It is always necessary for this type of network

c) It increases the chances of a global optimum being reached

Which of the following deep learning frameworks is not a library but an application programming interface that can run on top of various deep learning frameworks? a) Torch b) Theano c) Keras d) TensorFlow

c) Keras

Which of the following is the difference between knowledge-based systems and classical machine learning? a) Manually created representation b) Auto created features c) Mapping from features d) Simple features

c) Mapping from features

The most common simulation method for business decision problems that deal with probability in its characterization is called ________. a) Variable simulation b) Deterministic simulation c) Monte Carlo simulation d) Virtual simulation

c) Monte Carlo simulation

Which of the following is not among the topic modeling approaches? a) Latent Dirichlet allocation b) Latent semantic indexing c) Singular value decomposition d) Latent semantic analysis

c) Singular value decomposition

In the search engine development process, which of the following enablers are usually used to automatically read through the contents of Web sites? a) Classifiers b) Translators c) Spiders d) Regressors

c) Spiders

In sentiment analysis, which of the following is an implicit opinion? a) The hotel we stayed in was terrible b) Our new mayor is great for the city c) The customer service I got for my TV was laughable d) The cruise we went on last summer was a disaster

c) The customer service I got for my TV was laughable

What do voice of the market applications of sentiment analysis do? a) They examine employee sentiment in the organization b) They examine the stock market for trends c) They examine customer sentiment at the aggregate level d) They examine the "market of ideas" in politics

c) They examine customer sentiment at the aggregate level

Web site usability metrics include all of the following metrics, except? a) Page views b) Number of downloads c) User profiles d) Time spent on the site

c) User profiles

Which of the following statements is not true about Web site conversion statistics? a) Analyzing exit rates can tell you why visitors left your Web site b) The conversion rate is the number of people who act divided by the number of visitors c) Visitors who begin a purchase on Web sites must complete it d) Web site visitors can be classed as either new or returning

c) Visitors who begin a purchase on Web sites must complete it

________, a very popular NLP application, is trained on an enormous corpus of text data from books, websites, articles, and a variety of other human language sources to answer virtually any question. a) Google Translate b) Google Lens c) CNN d) ChatGPT

d) ChatGPT

________, often used interchangeably with AI, is the umbrella term used for technologies that rely on data and scientific methods/computations to make (or help/support in making) decisions. a) Computational linguistic b) Quantum computing c) CUDA d) Cognitive computing

d) Cognitive computing

Which of the following variable types in quantitative models are controlled and determined by the decision maker? a) Intermediate variables b) Stochastic variables c) Result variables d) Decision variables

d) Decision variables

As it is being practiced today, ________ seems to be nothing but an extension of neural networks capable of handling more complicated tasks with a higher level of sophistication. a) Rule based expert systems b) Fuzzy inference systems c) Support vector machines d) Deep learning

d) Deep learning

Which of the following methods calculates the values of the inputs necessary to achieve a desired level of output? a) Multiple regression b) Utility theory c) What-if analysis d) Goal seeking

d) Goal seeking

________ modeling uses rules to determine solutions that are good enough. a) Predictive b) Simulation c) Forecasting d) Heuristics

d) Heuristics

Understanding which keywords your users enter to reach your Web site through a search engine can help you understand ________. a) Most of your Web site visitors' wants and needs b) The hardware your Web site is running on c) The type of Web browser being used by your Web site visitors d) How well visitors understand your products

d) How well visitors understand your products

Managers in organizations typically have ________. a) A limited number of goals that can be independently optimized using linear and nonlinear programming b) Single goal that can be optimized using linear and nonlinear programming c) Single goal that cannot be optimized using linear and nonlinear programming d) Multiple goals that need to be simultaneously or jointly optimized

d) Multiple goals that need to be simultaneously or jointly optimized

Which of the following does not describe a deep feedforward network? a) Tensors are handled as input b) Most general type of deep network c) Many layers of neurons d) Network connections can be unidirectional or bidirectional

d) Network connections can be unidirectional or bidirectional

________ is among the most popular techniques proposed for shedding light into the "black-box" characterization of trained neural networks. a) Explanatory statistics b) Regression analysis c) Transparency analysis d) Sensitivity analysis

d) Sensitivity analysis

All of the following statements about data mining are true, except? a) The term is relatively new b) Intense, global competition make its application more important c) Its techniques have their roots in traditional statistical analysis and artificial intelligence d) The ideas behind it are relatively new

d) The ideas behind it are relatively new

All of the following statements about data mining are true, except? a) The novel aspect means that previously unknown patterns are discovered b) The valid aspect means that the discovered patterns should hold true on new data c) The potentially useful aspect means that results should lead to some business benefit d) The process aspect means that data mining should be a one-step analytics task

d) The process aspect means that data mining should be a one-step analytics task

Which of the following is not a concept related to information retrieval? a) Document matching b) Search engines c) Link analysis d) Topic modeling

d) Topic modeling