IS524 Final Outline

EXAMPLE TEST

PART II – Example short answer questions (10 x 4pts)

In general, for short answer questions, you must base your answer on frameworks discussed in class, perhaps extending/synthesizing or applying them to new situations in the process. Answers need to be relevant to class discussions/powerpoints/notes i.e. please do not pull answers “off the top of your heads”.

  1. In what way is a spreadsheet software an example of a Decsion Support System?

Dialog Management – the spreadsheet is the interface to enter the problem

Model Management – spreadsheet provides capability to create and link formulae

Data Management – spreadsheet has capabilities to select, filter and sort data

  1. Compare data mining capabilities with DBMS capabilities.

Data mining: What-if, sensitivity, goal seeking,optimization

DBMS capabilities: Those concerning data include querying and reporting. While a variety of reports are possible, there is no capability to analyze the data.

  1. A large cellular carrier wants to predict customers who might subscribe to its data plan. What techniques are suitable to identify such customers? Briefly explain.

Classification. This is a technique to group potential customers into two categories – those who might subscribe to its data plan and those who would not be interested. Classification can be carried out with Bayesian approach, neural net or discriminant analysis.

  1. Thessalonika Travels (TT) is an e-tailer that offers low cost vacation packages in the Mediterranean. The marketing manager at TT wants to know if the company’s existing banner advertising on Yahoo is an effective way of attracting prospective vacationers to the site. Based on analysis of site traffic, he/she has estimated that out of every 1,000 visitors to Yahoo, 50 of them visit the TT site. Out of this traffic (i.e. coming from the Yahoo site) 75% of the visitors are purchasers whereas in general 15% of all visitors for sites of this type are purchasers. Given that a visitor at TT is a buyer, what is the probability that the individual is someone from the banner site? Assume also that out of every 1,000 visitors to TT, 40 are from the banner site. In your view then, is it worthwhile to use banner advertising, given only the numbers and not costs?

P(buyer| from banner) = 0.75.

P(buyer) = 0.15.

P(banner) = 40/1000 = 0.04

P(from banner| buyer) = [P(Buyer| from banner) * P(from banner)]/P(buyer)

= [0.75 * 0.04]/0.15 = 0.20

This is compared to 0.15, the probability that someone will buy regardless of whether or not they are from the banner site. Its therefore worth while to go for banner advertising from the point of view of probability.

Note that these type of problems have to be clearly explained as shown above.

  1. What BI concepts are covered in the Continental Case Study? – identify and explain.

Dashboard – The dashboard could support drilldown, reporting, metrics (usage of metrics is implied, to measure airline success but not stated).

Warehouse – BI is updated from operational databases such as FSIR, passenger reservations

  1. Discuss the technical feasibility and time frame for an artificial eye that can see clearly for a distance of two miles (when the skies are clear and there are no obstructions).

n  Military already has satellites that can “see” much greater distances

n  Hardware is thus available; not clear whether image identification is included; assume it is not.

n  Sensory processing and image identification on a vast scale, in real time with miniaturization

will be a very imperfect process

n  For AI and miniaturization to develop 20-23 years possibly

(need to understand that it is a HW and a SW problem, HW challenges in miniaturization and SW in image recognition).

  1. Identify the representation scheme you will use for the following items of knowledge providing appropriate rationale and then model the knowledge using hypothetical examples:

a.  listing components and sub-components of an automobile engine and their functions

b.  lab policies (such as drives must be virus checked before being inserted etc.)

c.  providing physical details of a management training workshop (“how to communicate effectively”) that you attended in San Diego recently.

  1. listing components and sub-components of an automobile engine and their functions

semantic net – suitable for relationships. Also a declarative approach suitable for describing factual data such as about engines.

Note: the functions will be described with text nodes attached to each of the concepts for e.g. a valve opens and shuts to let the fuel in.

b. lab policies (such as drives must be virus checked before being inserted etc.)

rules because policies are a higher form of rules; they serve to guide/restrict actions in situations. But will be imperfect because policies will not lend themselves to if-then format.

If logging_in = true then do_virus_check = true.

If time = 8:00 pm and dayofweek > weekendday then close_lab = true.

If close_lab then shutdownmachines = true.

  1. providing physical details of a management training workshop (“how to communicate effectively”) that you attended in San Diego recently.

Frames; Frames are suitable for grouping and presenting knowledge as a unit. Slots for workshop location, date, fee, speaker etc.

  1. What might be some challenges posed by a KM initiative in an organization?

Easiest way is to focus on KM cycle:

1)  Define strategy – It is difficult to identify gaps in knowledge. Knowledge mapping is resource intensive, good tools do not exist.

2)  Identify sources – Experts may not exist in certain areas; need to be rewarded for sharing knowledge.

3)  Generate – Requires contact time; experts may be busy.

4)  Transfer/Codify – KM Systems and AI is not sufficiently advanced to store complex large scale knowledge

5)  Use/measure – Measurement of state of knowledge is subjective. Methods are still under development.

10.  A car company makes three types of models, sports cars, suvs and sedans. The profit margins on these are $5000, $1,000 and $600. The sports car requires 2,000 hours of labor while the SUV and sedan require 1,500 and 1,000 respectively. There is a production limit of 40,000 hours of labor. The sports car costs $15,000 in components while the suv and sedan cost $12,000 and $10,000 respectively. There is a budgetary restriction of $1,000,000 per month for components. The company needs to know how many of each type of car it should make each month in order to maximize its profits while fulfilling the constraints. Formulate an LP model.

Product / Hours / Components / Profit
SC / 2,000 / $15,000 / $5,000
SUV / 1,500 / $12,000 / $1,000
SDN / 1,000 / $10,000 / $600

Let SC -- # of sports cars; SUV – number of SUVs and SDN -- # of sedans to be manufactured

OF: 5,000 * SC + 1,000 * SUV + 600 * SDN maximize;

15,000 * SC + 12,000 * SUV + 10,000 * SDN <= 1000000;

2,000 * SC + 1,500 * SUV + 1,000 * SDN <= 40,000;

11.  What model would you use to show relationships between “stock market” (an index such as Dow Jones), “oil price”, “housing starts” (new houses built) and “economy”?

Causal Model because it can show causes and effects between these economic indicators.

12.  In Orange County, 51% of the adults are males. One adult is randomly selected for a survey involving credit card usage.

a. what is the probability that the selected person is a male?

b. It is later learned that the selected survey subject was smoking a cigar. Also, 9.5% of males smoke cigars, whereas 1.7% of females smoke cigars (based on data from the Substance Abuse and Mental Health Services Administration). Use this additional information to find the probability that the selected subject is a male.

(a)

P(male) = .51

(b)

P(male|smoking cigar) = [P(smoking cigar|male) * P(male)]/P(smoke cigar)

P(smoking cigar|male) = 0.095 (given)

P(male) = 0.51

P(smoke cigar) = #of cigar smokers/total population

{we are given that 9.5% of males i.e. 51% of population and 1.7% of females or 49% of population smokes cigars}

= 0.095*0.51 + 0.017*0.49 = 0.56

P(male|smoking cigar) = 0.095 * 0.51/0.56 = 0.85

12.  You have done a similar problem for quiz – will discuss in class.

13.  When faced with an important decision, a certain department manager was in the habit of going into the company’s retail outlets and talking to sales people/customers. The manager also would browse articles relating to the decision and would arrive at a decision without doing any formal analysis. What type of decision making is the manager following? Explain.

Intuitve decision making – this does not follow any set method; formal methods are especially not followed in this style of decision making.

5